Towards Photoinduced Charge Accumulation in

1 downloads 0 Views 5MB Size Report
von. Annabell G. Bonn aus Miltenberg, Bayern ...... In addition to undergoing IC, ISC, or multi phonon relaxation (MPR), the system can return to its ground state ...
Towards Photoinduced Charge Accumulation in Purely Molecular d6 Metal Complexes with Oligo-Triarylamine Donors

Inauguraldissertation

zur Erlangung der Würde eines Doktors der Philosophie

vorgelegt der Philosophisch-Naturwissenschaftlichen Fakultät der Universität Basel

von

Annabell G. Bonn aus Miltenberg, Bayern

Basel 2015

Genehmigt von der Philosophisch-Naturwissenschaftlichen Fakultät auf Antrag von

Fakultätsverantwortlicher/Dissertationsleiter: Prof. Dr. Oliver S. Wenger Korreferentin: Prof. Dr. Catherine E. Housecroft

Basel, den 24.03.2015

Prof. Dr. Jörg Schibler

Scientists will eventually stop flailing around with solar power and focus their efforts on harnessing the only truly unlimited source of energy on the planet: stupidity. Scott Adams

Für meine Eltern und Ella

Acknowledgement I thank Prof. Dr. Oliver S. Wenger for accepting me in his group and thus giving me the chance to work on such an interesting topic, I’ve developed a rather intense love-hate relationship with. I do thank him very much for his ever open door, the discussions, and his constant support. I couldn’t have wished for a better Doktorvater. I thank Prof. Dr. Catherine E. Housecroft for kindly agreeing to be my 2nd examiner. Sincere thanks is given to Prof. Dr. Thomas Pfohl for chairing my examination. I am very grateful to Sylvie Mittelheisser for measuring the EA, to Dr. Heinz Nadig for measuring all the ESI-HRMS spectra, and to Dr. Markus Neuburger for measuring and solving the OTA crystal structure. Very special thanks go to the ’unsung heroes’: The whole Werkstatt-Team for making the move to Basel and the every-day lab life as smooth as possible. Brigitte Howald and Beatrice Erismann for their help in all organizational matters. Special thanks goes to Markus Hauri for always being available and ever so helpful. The whole Häussinger group, particularly Heiko, Kaspar, and Thomas I thank for the occasional help with my NMR’s and more importantly for their readiness to participate in inter-group mingling and the coffee breaks. My thanks also goes to Angelo Lanzilotto for the introduction to the QuantaurusTau/-QY and, most importantly, for the key. Concerning the start up with LATEX, I thank Johannes Becker for not having me start completely from scratch. For the help in the synthesis of two reference molecules, I do thank my very last intern Benedikt Ganter.

I want to thank all past and present members of the AK Wenger so much: For the warm welcome in Göttingen, the joint efforts in moving to Switzerland, and the great time in Basel. Particularly, Martin for discussing (photo-)chemistry with me and answering all the questions, even the JACS paper couldn’t help me with. Ann-Christin for the pep-talks and our conversations about everything, but ’Gott und die Welt’. Julia for sharing a sink and bearing with me especially in the final stage of my presence in the lab. Laura, Hauke, and Andrea I thank very much for introducing me to a sport I, against all odds, really do enjoy a lot – Look, it’s a muscle! Last but not least, I thank my fellow in misery Luisa for sharing the burden with me and restoring my faith in blondes. A very special thanks goes to the “harte Kern” of the AK Wenger for being reliable companions on many an Apéro. I think you know who you are! Furthermore, I want to thank all the people I was fortunate enough to meet here in Basel, making the decision to move the right one. My aunt Jane and uncle Horst, I thank a lot for the provision of spatial, technical, sweet, and liquid support during the last four years. There are a handful of people who deserve special honors for helping me to stay on track these past years: Roland, I thank you very much for the years together and your encouragement and support during those last months. The most important people I couldn’t have done it without are my Ma, Pa, and my sister Vi. You have supported every decision I have made so far, especially moving to Basel, and will most likely still do so in the future. I do not have words that can express my deepest gratitude for your lifelong support and faith in me and my abilities.

Hypotheses Section III.1 Photoinduced hole accumulation on OTA in a Re2–OTA triad. Increased chance of absorption of a second photon due to the presence of two rhenium photosensitizers. Section III.2 Twofold charge separation resulting in AQ2––Re–OTA2+ after consecutive absorption of two photons. Increase of the first charge-separated state lifetime due to prolongation of the D–A distance and addition of Brønsted acids (PCET). Section III.4 Substitution of rhenium by ruthenium in AQ–Ru–OTA allowing for significantly higher excitation densities. Addition of metal ions (Lewis acids) as substitutes for stronger Brønsted acids to circumvent premature oxidation of OTA and promote the second charge separation step. Section III.6 Detection of an extremely long-living photoproduct in the time-resolved experiments in presence of Sc3+. Photo-accumulation of the second scandium ion-stabilized, charge-separated state AQ2––Ru–OTA2+ on UV-Vis timescale. Section III.7 Combination of two AQ acceptors and two ruthenium photosensitizers with one OTA donor in a pentad. Allowing for the accumulation of two holes on OTA entirely without the help of additives. Section IV Employment of the strong electron donor OTA in push-pull systems in combination with boron mesityl acceptors leads to superior CT and solvatochromic behavior.

Abstract One major problem humanity faces these days is the use of fossil fuels and even more pressing, the consequences of their combustion. Therefore, multiple research groups of various scientific backgrounds have made it their business to find a way to produce energy from sunlight and water. In order to successfully produce these so-called solar fuels, one has to go beyond charge transfer on a single-electron level. Most fuelforming reactions, such as H2 production or CO2 reduction, require multiple redox equivalents. The main focus of this thesis was on the creation of a simple, purely molecular (nanoparticle-free) donor–photosensitizer–acceptor (D–P–A) assembly capable of twofold photoinduced charge accumulation without the use of sacrificial agents. Anthraquinone (AQ) was chosen as the two-electron acceptor, on which charge built-up can be mitigated by proton-coupled or metal ion-coupled electron transfer. Oligotriarylamine (OTA) displays excellent ability to donate two electrons. In combination with d6 transition metal complexes, AQ and OTA present themselves as viable building blocks for D–P–A compounds for the study of photodriven charge accumulation. Initial photophysical investigation of rhenium containing triads gave charge separation on a single-electron level. In an AQ–Re–OTA triad, the charge-separated state lifetime could be significantly increased in presence of a strong Brønsted acid due to PCET. A long lifetime of the first charge-separated state is an important prerequisite for a second, charge-accumulating step to occur. Advancing the systems by incorporation of ruthenium polypyridyl photosensitizers allowed for higher excitation densities. AQ–Ru–OTA exhibited a lifetime in the microsecond range for the first charge-separated state already without acids. In presence of Sc3+, the lifetime was even increased beyond the detection limit of a nanosecond transient absorption setup. In the AQ–Ru–OTA triad, a ruthenium–OTA dyad, and in an AQ–Ru–OTA–Ru–AQ pentad, spectral evidence for OTA2+ was detected in presence of non-sacrificial additives, such as tetracyanobenzene or diphenylthiourea. Furthermore, push-pull systems comprised of OTA and a boron mesityl acceptor were studied with regard to their emission behavior and the suitability of OTA for use in charge transfer emitters. This lead to guidelines for the design of new charge transfer fluorophores.

Contents

I

General Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

II

Background and Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

II.1

Natural Photosynthesis and Water Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5

II.2

Fuel Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8

II.3

About The Excited State – Population, Fate, and Photochemistry . . . . . . . . . . 11

II.3.1

Population and Fate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

II.3.2

Photoinduced Electron Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

II.3.3

Energy Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

II.4

Accumulative Electron Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

II.5

Artificial Photosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

II.6

Chosen Molecular Moieties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

II.6.1

Metal Photosensitizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

II.6.2

Anthraquinone Acceptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

II.6.3

Oligo-triarylamine Donor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

III

Towards Photoinduced Charge Accumulation in d6 Metal Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

III.1 Triads with Two Rhenium Photosensitizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 III.1.1 Synthesis of the OTA Donor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 III.1.2 Synthesis of Precursors and the Reference Molecules . . . . . . . . . . . . . . . . . 31 III.1.3 Synthesis of the Studied Triads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 III.1.4 Characterization by IR and Optical Absorption Spectroscopy . . . . . . . . . . 34 III.1.5 Determination of the Redox Potentials and Establishment of the Energy Level Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 III.1.6 Steady State Luminescence Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 III.1.7 Possibility of Triplet-Triplet Energy Transfer . . . . . . . . . . . . . . . . . . . . . . . . . 40 III.1.8 Time-Resolved Luminescence Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . 41

I

Contents III.1.9 Identification of the Photoproducts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 III.1.10 Oxidative Flash-Quench Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 III.1.11 Influence of the Excitation Density on the Photochemistry . . . . . . . . . . . . 47 III.2 Electron Transfer in an AQ–Re–OTA Triad . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 III.2.1 Synthesis of AQ–Re–OTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 III.2.2 Spectroscopical Characterization of the Triad . . . . . . . . . . . . . . . . . . . . . . . . 54 III.2.3 Electrochemical Study of AQ–Re–OTA and Establishment of the Energy-Level Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 III.2.4 Influence of Acids on the OTA Donor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 III.2.5 Influence of Hydrogen-Bond Strength on the Photoproducts . . . . . . . . . . 58 III.2.6 Kinetics of the Photogenerated Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 III.2.7 Dependence of the Transient Spectral Absorption on the Excitation Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 III.3 Summary and Conclusions on Photoinduced Electron Transfer in Rhenium-OTA-Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 III.4 Photoinduced Electron Transfer in an AQ–Ru–OTA Triad . . . . . . . . . . . . . . . . . 64 III.4.1 Synthesis of the AQ–Ru–OTA Triad and a Reference Dyad . . . . . . . . . . . . 64 III.4.2 Optical Spectroscopic and Electrochemical Characterization . . . . . . . . . . . 67 III.4.3 Kinetic Emission Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 III.4.4 Chemical Oxidation of OTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 III.4.5 Time-Resolved Absorption Studies on the Ru–OTA Dyad . . . . . . . . . . . . . 71 III.4.6 Time-Resolved Study of AQ–Ru–OTA in Pure MeCN . . . . . . . . . . . . . . . . . 74 III.4.7 Time-Resolved Study of AQ–Ru–OTA in Presence of Lewis Acids . . . . . 75 III.4.8 Time-Resolved Studies on AQ–Ru–OTA in presence of Sc(hmpa)33+ . . . . 83 III.4.9 Time-Resolved Studies in Presence of N,N-Diphenylthiourea . . . . . . . . . . 84 III.4.10 Summary on Electron Transfer in AQ–Ru–OTA . . . . . . . . . . . . . . . . . . . . . . 87 III.5 Ferrioxalate Actinometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 III.6 Continuous Irradiation Experiments on Ru–OTA and AQ–Ru–OTA . . . . . . . . 93 III.6.1 Reference Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 III.6.2 Irradiation of Aerated Samples in Presence of Sc3+ . . . . . . . . . . . . . . . . . . . 96 III.6.3 Irradiation of Aerated Samples in Absence of Sc3+ . . . . . . . . . . . . . . . . . . . . 100 III.6.4 Dark Reactions of Aerated Samples in Presence and Absence of Sc3+ . . 101 III.6.5 Irradiation of De-Oxygenated Samples in Presence of Sc3+ . . . . . . . . . . . . 102 III.6.6 Irradiation of De-Oxygenated Samples in Absence of Sc3+ . . . . . . . . . . . . 104

II

Contents III.6.7 Dark Reactions of De-Oxygenated Samples in Presence and Absence of Sc3+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 III.6.8 Possibility of Photo-Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 III.7 Towards Accumulative Electron Transfer in an AQ–Ru–OTA–Ru–AQ Pentad108 III.7.1 Synthetic Pathway to AQ–Ru–OTA–Ru–AQ . . . . . . . . . . . . . . . . . . . . . . . . . . 108 III.7.2 Electrochemical Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 III.7.3 Optical Spectroscopic Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 III.7.4 Photophysical Behavior in Different Solvents . . . . . . . . . . . . . . . . . . . . . . . . . 112 III.7.5 Time-Resolved Experiments in Different Solvents . . . . . . . . . . . . . . . . . . . . 113 III.7.6 Photophysical Behavior in Presence of Thiourea . . . . . . . . . . . . . . . . . . . . . . 119 III.7.7 Summary and Conclusion on Electron Transfer in AQ–Ru–OTA–Ru–AQ122

IV

Solvent Dependent Emission of OTA–B(mes)2 Molecules . . . 125

IV.1 Synthesis of Studied Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 IV.2 Solvent-Dependent Charge Transfer Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 IV.2.1

Electrochemical Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

IV.2.2

Absorption and Emission Behavior in Hexane . . . . . . . . . . . . . . . . . . . . . . . 130

IV.2.3

Solvatochromism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

IV.2.4

Difference between Ground and Excited State Dipole Moments . . . . . . . 134

IV.2.5

Luminescence Lifetimes and Quantum Yields . . . . . . . . . . . . . . . . . . . . . . . . 137

IV.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

V

General Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

VI

Experimental Part . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

VI.1 Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 VI.2 Syntheses of the OTA donor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 VI.3 Syntheses of General Building Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 VI.4 Syntheses of ligands and metal complexes with rhenium . . . . . . . . . . . . . . . . . . 151 VI.5 Syntheses of ligands and metal complexes with ruthenium . . . . . . . . . . . . . . . . 159 VI.6 Syntheses of the D–bridge–A molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

VII Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

III

Contents

A

Abbreviations and Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V

B

List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX

C

List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XIX

D

Curriculum Vitae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXI

IV

I. General Introduction

With world’s population growing, estimated to reach 9.6 billion by 2050,1 humanity needs to find a way to secure Earth as a living environment for as long as possible. The number of inhabitants is going to increase significantly and the energy consumption per head and thus the absolute energy demand has already increased measurably. With the rapid technological progress and the power this continues to require, mankind’s energy needs will most likely increase at a much steeper rate than population itself. According to estimates, the world’s energy consumption rate will have increased to 40.8 TW by 2050. For comparison, in 2001 it was 13.5 TW.2 So far, fossil fuels like coal, crude oil, and gas have been the means of choice to meet the energy needs. In 2013, 85% of our energy requirements have been covered by fossil fuels. Thus, they have enabled new inventions and further advancement. Though abundant at times, only limited reserves are left and most of them are classified as ‘not recoverable’. However, our immediate problem does not lie in the lack of fossil fuels but in the consequences of their combustion. Having burned them for years, we continuously polluted Earth with greenhouse gases (GHG) such as CO2 or the even more harmful, synthetic hydrofluorocarbons (HFCs). The latter are mostly industrially released, which is again driven by fossil fuel combustion. Therefore, we promoted the depletion of the stratospheric ozone layer and global warming. Fossil fuels are essentially biomass generated via natural photosynthesis millions of years ago. Their combustion is simply reverse photosynthesis, setting free again the pollutants Nature so elegantly recycled. We tend to call the biomass we burn today, and which covers about 11% of our energy needs, renewable energy.3 This denotation is certainly not quite accurate, if there are no plans afoot to e.g. reforest the woods we lumber. The truly renewable energy source we should focus on is solar energy. In the beginning of the 20th century, scientists realized that light, if converted properly, can serve us as an infinite energy source.4 With 100000 TW per annum, one hour of sunlight could cover the energy needs of mankind for one year. Another abundant material on earth is water. Direct dissociation of H2O into hydrogen and oxygen by simple irradiation with sunlight, however, is not possible. The

1

I General Introduction lack of overlap between the absorption spectrum of H2O and the emission spectrum of sunlight prevents this.5 If we manage to find a way for using sunlight to split water into hydrogen and oxygen, we will be close to solving our energy problems. Both resources are present in abundance, hydrogen gas is considered the most environmentally friendly fuel, and oxygen would be the only side-product. Since the emergence of the field of photochemistry, interdisciplinary research teams have devoted themselves to understanding the working principles of natural photosynthesis. With that, Nature has provided a blue-print for successful solar energy conversion and fuel production. It is certainly not desirable to build a faithful replica of PSII to cover our energy needs. For once, it will most likely be synthetically very demanding, if not unfeasible at all. Furthermore, natural photosynthesis in plants and bacteria does not exhibit a high quantum yield for energy conversion. It was designed by Nature as a repetitiously successful process with a high turn-over number. As a result, the beauty of energy harvesting, conversion and storage in the photosynthetic centers does not lie in their efficiency. Around 80% of the initially absorbed light energy is discarded again, unused, because it simply cannot be processed by the photosynthetic organism. Natural photosynthesis was developed for the sake of survival, not for efficient, excessive energy production the organism has no need for.6,7 Nowadays, we have a very firm grasp on the underlying mechanisms of the photosynthetic process. We know that charge separation, creating a high energy state, plays a pivotal role. More accurately, the separation of multiple redox equivalents is crucial for the production of high-energy substances, i.e. solar fuels. So far, we have made use of sunlight in various types of manufactured solar cells.3 They provide the necessary electricity for driving electrochemical water splitting, producing oxygen and hydrogen. The latter can then directly be used as fuel or converted into other high-energy compounds, e.g. carbohydrates. Another possibility for hydrogen production is the use of photosynthetic organisms such as cyanobacteria or green algae.8 Unfortunately, though indeed producing solar fuel, the above mentioned ways do so in an indirect fashion. Losses are inevitable. The bio organisms, for example, will use a certain amount of the solar energy to survive. Therefore, designing a single system that combines the sites for direct light harvesting, solar energy conversion by charge separation, and fuel generation is preferable. By successfully mimicking the essential steps in natural photosynthetic centers, we might one day achieve fully renewable fuel production from sunlight and water. Figure I.1 depicts the schematic structure of such a proposed system. An important prerequisite for the conversion of H2O and also CO2 to usable fuels is the accumulation of multiple redox equivalents. For this, a profound knowledge

2

I General Introduction hn O2 evolving catalyst

ET

D

ET

P

ET

A

ET

H2 evolving catalyst

Figure I.1 Schematic structure of a system combining light-harvesting, charge separation, and fuel production. D = donor, P = photosensitizer, A = acceptor, ET = electron transfer.

of the factors governing successful charge accumulation is essential for the creation of a working example of the system in Figure I.1. Reported systems so far, have achieved only either water oxidation or fuel production and they are usually in need of high-energy sacrificial (“dead-end”) reactants. However, the ultimate artificial photosynthetic system has not yet been realized. The main aim of this thesis was on achieving charge accumulation in purely molecular, nanoparticle-free D–P–A assemblies. They were designed to combine the half-reactions of hole and electron accumulation on two specific sites within one single molecular system. The focus hereby was on the realization of a self-sufficient central building block D–P–A as depicted in Figure I.1, which would not rely on sacrificial agents anymore. For this, purely molecular (nanoparticle-free) systems theoretically capable of twofold charge accumulation via successive photoinduced electron transfer were designed, synthesized, and investigated physicochemically. Rhenium or ruthenium polypyridyl complexes were employed as photosensitizers. Anthraquinone (AQ), able to accept up to two electrons, was used as the acceptor in adaption of the quinone acceptors present in natural photosynthesis. The role of the electron donor was in all cases adopted by an oligo-triarylamine (OTA). This very strong donor can sequentially release up to three electrons. The manufactured systems were usually comprised of these three main molecular moieties covalently linked and combined in various combinations. By always improving the complex setup it was worked towards achieving twofold charge accumulation on the donor and/or acceptor site without the help of sacrificial reagents. The following chapter will first give an introduction into natural photosynthesis and the oxygen evolving catalyst. A profound understanding of the systems we are trying to mimic in order to generate solar fuels is essential. Subsequently, the basic principles of photochemistry are introduced, as well as the problems associated with the accumulation of multiple redox equivalents. This will be followed by an overview

3

I General Introduction of reported charge accumulating systems in literature and the molecular units chosen in this thesis. In a small side-project unrelated to artificial photosynthesis, the oligo-triarylamine donor was employed in a series of organic D–A dyads. These push-pull systems were investigated with respect to their solvent-dependent charge transfer properties. A lot of research is still focusing on the creation of very powerful, versatile emitters, and efficient blue and pure white light emitting compounds for use in OLEDs still attract a lot of scientific attention.9,10

4

II. Background and Motivation

II.1 Natural Photosynthesis and Water Oxidation Over millions of years, Nature has established and perfected a means to capture sunlight and store the solar energy by converting CO2 and H2O to carbohydrates and other high-energy substances – the photosynthetic process. This endeavor requires the fine-tuning of several complex processes: light-harvesting by antenna chromophores, charge separation by several consecutive electron transfer steps, regeneration of the reaction center by oxidation of water, and CO2 conversion to high-energy compounds, such as sugars. The photosynthetic process is usually separated into two individual steps, the light reaction and the dark reaction. In the latter, CO2 is captured and the solar energy acquired in the light reaction is ultimately stored in form of high-energy carbohydrates which then serve as ‘fuel’ for the organism. The focus of the following section, however, will be on the light-dependent reactions taking place in photosystems I and II (PSI and PSII) located in the thylakoid membranes of the chloroplasts. This is also where the oxidation of H2O occurs. To mimic this reaction is the primary objective of many research groups.

P700*

A1

NADPH

FD

potential energy

FNR

P680* 2 H 2O

O2 + 4H

PQ

Cyt f

NADP+ + H+

PC

P700 OEC

+

Pheo

TyrZ

PS II P680 PS I

Scheme II.1 Simplified Z-scheme of the light-dependent reactions in PSI and PSII.

Scheme II.1 depicts the charge-transfer chain as followed in the photosynthetic process. This simplified representation is referred to as a so-called Z-scheme.11 The first

5

II Background and Motivation step of the photosynthetic process is the capture of solar energy by specially designed antenna chromophores. These transfer the excitation energy to P680, a chlorophyll dimer located in the reaction center of PSII, within 1 ps. The photoexcited P*680 then reduces a nearby pheophytin (Pheo), the primary electron acceptor. As a result a + primary charge-separated state, P680 Pheo–, is formed. This initial charge-separating step exhibits a thermodynamic efficiency of 70%.12 The electron is then shuttled along an electron transfer chain with two protein bound quinone acceptors QA and QB (not shown in Scheme II.1), located in the plastoquinone pool (PQ). Upon two successive excitation and charge-separation steps the twofold reduced QBH2 leaves PSII via the PQ on the other side of the protein. The reducing equivalents are then transported further along another electron transfer chain over cytochrome f (Cyt f), plastocyanin (PC), phylloquinone (A1), and a ferredoxin unit (FD) among others. With the excitation of P700 enough energy is accumulated to fixate CO2 at a ferredoxin-NADP reductase (FNR) where NADP+ and 2 H+ are reduced to produce NADPH, the natural form of H2 storage. CO2 can then be transformed into high-energy substrates by means of NADPH. To ensure efficient use of the harvested solar energy, the rapid spatial separation of both, electrons and holes is crucial. Within 200 ps the electron is separated from the + excited P*680 by 26 Å to form P680 Q–. This fast initial electron transfer enables a stable charge-separated state with a lifetime of hundreds of microseconds.6 The oxidized + chlorophyll P680 , as a strong oxidant (Eox = +1.2 V vs. NHE),13 is regenerated by electron transfer from a tyrosine unit (TyrZ).

Figure II.1 Left: Reaction center of PSII. The surrounding protein structure was omitted for clarity. Electron transfer pathways are indicated by arrows.14 Right: Crystal structure of the OEC with its ligand environment.15 Reproduced with permission from14,15 © Royal Society of Chemistry and Nature Publishing Group.

6

II Background and Motivation This electron is originally released by oxidation of the so-called oxygen evolving complex (OEC), a pentanuclear CaMn4-cluster located in the heart of PSII. In the last years, a lot of attention was given to the elucidation of the structure of OEC. Only quite recently Umena et al. were able to obtain a crystal structure of the very complicated oxygen-evolving PSII with a resolution as high as 1.9 Å.15–17 Figure II.1 displays the reaction center of PSII with the corresponding electron transfer pathways (left part). The right side depicts the OEC with its ligand environment. Comprised of one calcium atom and four oxygen-bridged manganese atoms, the OEC is also surrounded by a protein matrix separating it from the thylakoid lumen. (Figure II.1) Due to the surrounding amino acids the OEC is well insulated, allowing for a slow, stepwise oxidation. Another advantage of this insulation is the prevention of premature proton loss or uptake. As a result, the OEC is not converted into its reduced resting state prior to the release of O2 by complete oxidation of H2O.13,18

Scheme II.2 Extended S-state cycle according to Dau and Haumann, depicting sequential proton and electron removal from the CaMn4-cluster. The times correspond to the half-lifes of the S-transitions.18

The general mechanism of H2O oxidation, taking place on the OEC as suggested by Kok in 1970 proved correct. Since then, several groups attempted to elucidate the exact underlying mechanism. Kok’s initial catalytic cycle did not involve protons or reactions with H2O in any of the four steps. It only stated the stepwise accumulation of four oxidation equivalents on the OEC, as indicated by the different numbers of the S-states. Eventually, this would lead to the oxidation of H2O. Recent studies were able to shed light on a possible reaction mechanism involving H2O molecules bound in the vicinity of the CaMn4-cluster. Haumann et al. have experimentally studied pigments of PSII by time-resolved EPR, UV-Vis and XAS spectroscopy and summarized their results in the extended S-cycle as depicted in Scheme II.2.18 The given times correspond to the half-lifes of the S-transitions. A full cycle requires the sequential absorption of four photons. Due to dark-adaption, S1 is the predominant, long-term dark-stable

7

II Background and Motivation state. Therefore, the starting point was found to be an S1 S2 transition. Upon the first photon uptake, oxidation of one manganese atom is induced. DeprotonaS3 tion, however, was not observed. A second photon uptake, promoting an S2 + transition leads to the formation of TyrZ with concomitant manganese oxidation and proton release. The absorption of a third photon starts a series of transitions, S3 S4 S4’ S0, which ultimately result in O2 formation by H2O oxidation. The S3 S4 transition is believed to involve proton release. Whereas the subsequent electron transfer from manganese to TyrZ takes place in the S4 S4’ transition. In the next step, H2O is oxidized, O2 released, and manganese oxidized one last time. After proton release the OEC reaches S0. With a final S0 S1 transition the cycle is closed. No charge accumulation occurs in this last step, since the electron transfer is supposed to be accompanied by deprotonation.19 In addition to the extended S-cycle in Scheme II.2, Dau and Haumann have proposed a cycle of alternating proton and electron release which would entail more intermediates still in need of characterization.18 Despite the well-resolved crystal structures of the OEC, the exact mechanics of water splitting is still partly speculative. The concomitant release of four protons and electrons to prevent the built-up of charge, however, is certain.

II.2 Fuel Production Nowadays, humanity aims to store solar energy in form of chemical bonds, producing so-called solar fuels. As shown by natural photosynthesis, these reactions entail multiple charge-separating steps. Water oxidation, reduction of CO2, and hydrogen production each require more than one electron to be transferred in order to make and break chemical bonds. Hydrogen Production The storage of energy in form of hydrogen gas is a very clean and environmentally friendly way, since its only oxidation product is H2O. Hydrogen could also be used directly as a fuel to drive machinery or be converted into electricity by means of fuel cells. Regarding the number of charges in need of accumulation, the production of H2 is one of the easier reactions to master. As shown in Equation II.1, only two electrons have to be accumulated which then react with two protons to form H2. 2 H+ + 2 e–

H2

(II.1)

As a result, a lot of research has focused on visible light-driven hydrogen generation.

8

II Background and Motivation One approach is sunlight-induced reforming of biofuels such as glucose, ethanol or methanol. For that, we make use of Nature by applying natural systems that already successfully produce hydrogen. These species are e.g. natural hydrogenase enzymes.8,20 The earliest purely man-made approach to homogeneous hydrogen production were molecular three-component systems. They were comprised of a metal photosensitizer, sacrificial reagents to re-reduce the photosensitizer, and an electron relay station. As the actual hydrogen producing site mostly colloidal platinum was employed. In recent years the concept was developed further, linking the light-harvesting photosensitizer and the photocatalyst covalently. In 1983 Lehn and co-workers reported a combination of Ru(bpy)32+ and a cobalt catalyst, which produced hydrogen in presence of a sacrificial electron donor. This system was among the first to directly reduce the photocatalyst by the photosensitizer.20–22 Water Oxidation The idea of using earth-abundant materials for future fuel production, inevitably leads to H2O as a potent basic raw material. Directly observable from the chemical equation for the splitting of H2O (Equation II.2) is that in addition to the generation of O2, protons and electrons are released. The latter can be recombined to yield H2.

2 H 2O

O2 + 4 H+ + 4 e–

1.23 V

(II.2)

The above reaction has a reported standard potential of 1.23 V vs. NHE. In reality, driving this reaction requires a higher potential, the so-called overpotential. It is needed to overcome activation barriers that are present for the electrolysis of pure water. As a result, to ensure efficient water splitting a minimum potential of ≈2 V is necessary.23 In Nature, PSII absorbs photons with a wavelength of 680 nm to drive the photosynthetic process and thus, water splitting in the OEC. One such photon carries 1.83 eV of energy. Therefore, also visible light should enable photocatalytic water splitting, despite the need for an overpotential. Nevertheless, water oxidation catalysts, as for example the OEC in PSII, are used to minimize the overpotential in order to facilitate the process. In 1972, Fujishima and Honda were the first to demonstrate heterogeneous photocatalytic water splitting in a photoelectrochemical cell by irradiation of TiO2 with UV light. Oxygen evolution occurred at the TiO2 photoanode, whereas hydrogen was detected at a platinum counter electrode.24 In contrast to heterogeneous photocatalysis as performed by Fujishima, the advantage of molecular systems is the easy customization of their structural and electronic

9

II Background and Motivation properties. Furthermore, their well-defined structure facilitates the study of underlying mechanisms. Meyer and co-workers were the first to report a molecular water oxidation catalyst, a dinuclear ruthenium polypyridyl complex (“blue dimer”, cis,cis-[Ru(bpy)2(H2O)]2(µ O)4+) in the 1980s.25 Following this pioneer work, multiple binuclear and mononuclear molecular water oxidation catalysts, mostly ruthenium polypyridyl complexes, have been reported in recent years. Yet, molecular complexes for water oxidation in homogeneous systems are still quite rare, compared to the heterogenous approach.23,26–29 CO2 Reduction Even more challenging than water oxidation, from a kinetic and thermodynamic viewpoint, is the photoelectrochemical reduction of CO2 to carbon monoxide or hydrocarbons as shown in Equations II.3. Preparation of hydrocarbons, e.g. formic acid, methanol or sugars is preferable. They are high-energy compounds that can be of direct use again. Whereas CO gas has to be further transformed, mostly industrially, into useful chemicals such as acetic acid, formic acid, or methanol. In all reactions listed below more than one reducing equivalent has to be accumulated. The potentials are given in V vs SCE at pH7.30 CO2 + 2 H+ + 2 e–

HCOOH

−0.85 V

CO2 + 4 H+ + 4 e–

HCHO + H2O

−0.72 V

CO2 + 6 H+ + 6 e–

CH3OH + H2O

−0.62 V

CO2 + 8 H+ + 8 e–

CH4 + 2 H2O

−0.48 V

(II.3)

Homogeneous photocatalysis for CO2 reduction usually makes use of a light absorbing molecule and a transition metal catalyst, both promoting proton-coupled transfer of multiple electrons. Single electron reduction of CO2 to form CO2 – requires a reduction potential of ≈ −2.14 V vs. SCE. Additionally, an overpotential of up to 0.6 V has to be considered.30 Therefore, the concomitant transfer of both electrons and protons is preferred. The latter help avoid charge built-up. One advantage of transition metal complexes is the tunability of their photophysical properties simply by ligand modification. One example of photoinduced CO2 reduction systems are cobalt or nickel tetraaza-macrocyclic complexes with [Ru(bpy)3]2+ as photosensitizer. In another approach, ruthenium polypyridyl sensitizers were covalently linked to Re(bpy)(CO)3X catalysts.30–32 In order to successfully perform the above listed reactions driven by light, a deeper

10

II Background and Motivation understanding of the mechanics, challenges, and problems associated with the accumulation of multiple redox equivalents is crucial.

II.3 About The Excited State – Population, Fate, and Photochemistry The following section will give an introduction in the basics of photophysics and photochemistry to the extent necessary to understand the topics and results of this thesis. It will mainly focus on the underlying concepts of electronic absorption of light and the possible deactivation pathways of the excited state. Deactivation can hereby occur via radiative or non-radiative pathways, electron or energy transfer.

II.3.1 Population and Fate In classical electrodynamics light is defined as high-frequency electromagnetic waves bearing a variety of different energies. However, to understand the concept of absorption and emission of light by matter, this approach is insufficient. Excitation of a transition metal photosensitizer P to form P* is initiated by the elevation of an electron from a low-lying energy state to a state higher in energy. Hereby, the minimum energy needed equals the difference between the energy levels of the ground state and excited state. The energetic position of the states is dependent on the nature of the complex. As a result, the energy needed can be quantified. Quantum mechanics, therefore, more accurately describes the transition of an electron as the absorption of a so-called photon with a defined energy content. From the absorption spectra of metal complexes one can deduce the probability of the different transitions. Spin-allowed tranπ* or 1 MLCT (metal-to-ligand charge transfer) sitions, such as ligand-centered π transitions produce absorption bands of rather high intensity. On the other hand, spin-forbidden transitions (3 MLCT, d-d transition) produce very weak bands and are more difficult to detect in absorption spectroscopy.33 The Jablonski diagram (Figure II.2) displays the different possible transitions upon absorption of a photon in a metal complex. All non-radiative pathways are shown by black, wiggly arrows. In the ruthenium and rhenium transition metal complexes used in this thesis, excitation occurs from a singlet ground state (S0) to a thermally excited singlet state (Sn). Faster than any photochemical or photophysical process, Sn relaxes to the vibrational ground state of the lowest-lying singlet excited state via radiationless deactivation (internal conversion IC, > 1011 s−1 ). Despite the excitation to a spin-allowed singlet

11

II Background and Motivation E Sn IC

S1 hn

Tn

ISC

T1

hnF MPR

Photoreactions

hnP

S0 Figure II.2 Simplified Jablonski diagram.

state, the actual luminescent, lowest-energy excited state in transition metal complexes is usually a triplet state T1. This state is accessible via intersystem crossing (ISC, 106 − 1011 s−1 ) from the S1 state. The transition between states of different spins is enabled by the strong spin-orbit coupling present in the transition metal complexes. In addition to undergoing IC, ISC, or multi phonon relaxation (MPR), the system can return to its ground state by releasing the excess energy as a photon. If the emission of light originates from S1 the radiative process is called fluorescence (hνF ). Whereas the spin-forbidden transition from T1 to the singlet ground state is referred to as phosphorescence (hνP ). It usually occurs at a much longer timescale than fluorescence exceeding the duration of excitation. The radiative deactivation processes are indicated by the solid arrows in Figure II.2.34,35 Once the photoluminescent state T1 of the metal complex is populated via the above discussed pathways it can be deactivated by various processes. Already mentioned were radiative deactivation by luminescence (hνF , hνP ) and radiationless deactivation by heat dissipation (IC, ISC). For the calculation of the excited state lifetime all possible radiative and non-radiative processes have to be considered. Each one of them is assigned a rate constant (ki) and the inverse of their sum gives the lifetime τ (Equation II.4 left). τ=

1 ∑i k i

φ=

nhν (emitted) nhν ( absorbed)

(II.4)

In addition to the lifetime of the excited state, its luminescence quantum yield φ is of interest. It can be calculated by dividing the amount of emitted photons by the number of absorbed photons (Equation II.4 right). In case of sufficiently long lifetimes of P*, it can partake in photochemical reactions forming new products, or it can be quenched by energy or electron transfer to or from nearby molecules (Equations II.5).36

12

II Background and Motivation

P* + A

P+ + A–

P* + D

P– + D+

P* + Q

P + Q*

(II.5)

If an electron donor is in close proximity to the excited photosensitizer, it can transfer an electron to P* quenching it reductively. In presence of an electron acceptor this process is called oxidative quenching. Furthermore, energy transfer between P* and a suitable quencher Q is possible.

II.3.2 Photoinduced Electron Transfer The simplest setup capable of undergoing electron transfer requires a donor D and an acceptor A. In the work presented in this thesis D and A are covalently linked via bridging units and in some cases the role of A is played by a photosensitizer. In the following, electron transfer theory according to Marcus and Hush will be introduced to the extent necessary for the understanding of this thesis.37 The original, classical theory about electron transfer as pioneered by R. A. Marcus addresses outer-sphere, self-exchange reactions. Marcus communicated his theory in numerous papers in extraordinary mathematical detail.38–44 Later, N. S. Hush extended the theory, taking inner-sphere reactions into account.45 In outer-sphere reactions, a term adapted from coordination chemistry, no chemical bonds are formed or broken, however, due to the transfer of electrons and the accompanying redistribution of charge, bond lengths and angles can change. Furthermore, the surrounding solvent shell, as well as the involved redox sites have to adapt to the changed charge distribution. Thus, the need for reorganization results in an increase of the activation barrier, which in turn decreases the reaction rate. Figure II.3 depicts the potential wells for an exergonic electron transfer reaction, as defined by Marcus-Hush theory. For non-adiabatic electron transfer reactions, i.e. weak electronic coupling (HAB) between D and A, the reactant and product states retain their individuality (cyan parabola). In this case, the electron has to hop from the reactant state (D–A) to the product state (D+–A–), overcoming the free energy of activation (EA). The energy of the activation barrier can be derived from the intercept of the two cyan parabolas. The horizontal axis represents the reaction coordinate and the vertical axis the free 0 energy for electron transfer GET . In Marcus-Hush theory, the reaction coordinate is a

13

II Background and Motivation

D-A +

D -A

-

l 2HAB EA 0

DG

Figure II.3 Potential energy wells for exergonic electron transfer in D–A including inner and outer sphere reorganization.

combined representation of changes in the outer-sphere (solvent shell reorganization) and inner-sphere (bond lengths and angles adjustment). The systems investigated in this thesis usually undergo non-adiabatic electron transfer, due to the relatively weak electronic coupling between the donor and acceptor units. For adiabatic electron transfer processes HAB is strong, resulting in a splitting of the energy wells at their intercept. The maximum splitting is 2 HAB. In this particular case, the charge transfer system remains on the lower potential energy surface when passing from D–A to D+–A– (blue parabola). For non-adiabatic electron transfer reactions the following semi-classical equation can be applied to calculate the rate constant. s k ET =

"  # 0+λ 2 ∆G 4π 3 H 2 · exp − 4λk B T h2 λk B T AB

(II.6)

Equation II.6 shows that the electron transfer rate is mainly dependent on three parameters: 1) the reorganization energy λ; 2) the reaction free energy, or driving force ∆ G0; and 3) the electronic coupling, HAB, between donor and acceptor. Further parameters needed to calculate the reaction rate are the Boltzmann constant kB and the temperature T.

Reorganization Energy

In order for electron transfer to occur, the reactants have to obtain nuclear configurations similar to the products. The necessary change in configuration is achieved by vibrational reorganization. When estimating the systems reorganization energy λ, contributions of both the inner and outer sphere have to be considered. For the inner sphere (Equation II.7) this means the adjustment of bond lengths and angles upon

14

II Background and Motivation charge transfer. Following a changed charge distribution in the D–A system, D and A have to adapt to the new situation. In Equation II.7, ki are the normal mode vibrational force constants and ∆qi corresponds to the changes in the nuclear configuration of the normal vibrational modes.   1 2 k ∆q 2∑ i i    1 ∆e2 1 1 1 1 = − + − 4πe0 eop es 2r D 2r A d DA λin =

λout

(II.7) (II.8)

In addition to changes in the inner-sphere of D–A, a transfer of charge always entails changes in the solvent shell as well. Directly observable from Equation II.8 is the dependence of the outer reorganization energy on the D–A distance dDA and the radii of D and A (rD, rA), respectively. For an easy estimation of the radii, D and A are regarded as hard spheres. The bigger these spheres, the smaller is the interaction between the solvent molecule and the charge on D or A, resulting in a small λout . Furthermore, the amount of transferred charge ∆e has to be considered. Most importantly, the value of the outer reorganization energy is dependent on the static (es ) and optical (eop ) dielectric constants of the surrounding solvent. The effect of solvent reorientation is especially prominent when charge transfer reactions occur in polar solvents. In the initial state D–A, usually a charge-neutral state, the electric field is weaker than in the charge-separated product state. Therefore, solvents with a strong electric dipole moment show a higher affinity to the product state. As a result, the contribution of λout to the total reorganization energy is quite small in apolar solvents.

Driving Force

When passing from the relaxed ground state of D–A to its charge separated D+–A– state, a certain amount of change in the free energy is observed. This change in energy is called the driving force of the electron transfer process, ∆G0. The transfer of charge from the reactant to the product energy well requires the system to overcome the activation barrier EA. For the calculation of this activation energy barrier, Equation II.9 can be applied. Semi-classical treatment of electron transfer processes has shown that EA mainly arises from the necessary reorganization of the inner and outer sphere due to the changed charge distribution in the reactant and product states. Therefore, λ is also part of Equation II.9. ∆G0 + λ EA = 4λ

2 (II.9)

15

II Background and Motivation The electron transfer rate constant, kET, reaches a maximum when −∆G0=λ. At this saddle point of the parabola in the lower part of Figure II.4, the system has reached its optimal electron transfer rate and no activation energy barrier is present. An increase of the reaction rate with increasing driving force (−∆G0) is a well expected and accepted phenomenon. 0

-DG > l

-DG = l

EA l

0

0

-DG < l

EA

EA = 0 DG l

DG

DG l

lg kET

normal region

inverted region -DG0

Figure II.4 Depiction of the normal and inverted region as proposed by Marcus.

However, given the quadratic contribution of ∆G0, Equation II.9 implies a decrease in the electron transfer rate with a further increase in exergonity. This counter-intuitive behavior starts when −∆G0>λ, as can be seen in the depiction of the different Marcus regions in Figure II.4. Once the saddle point is overcome, the system operates in the Marcus-inverted region. The phenomenon of the inverted region stems from the fact that in a system of very high exergonity, a lot of energy is released upon charge transfer. This much energy cannot be discharged by the system fast enough, which might lead to the proposed decrease in the reaction rate. Roughly 20 years after Marcus first postulated his theory, Miller and Closs have found experimental proof of the inverted region.46,47 They studied intramolecular electron transfer rates in a series of covalently linked D and A molecules. A rigid spacer allowed them to keep D and A at a fixed distance and variation of the driving force was achieved by altering D and A, respectively. A rough estimation of the driving force can be achieved by using the approximation ∆G0 ≈ e(Eox-Ered)-E00. Where E is the electronic ground state potential of the donor and acceptor, respectively, and e is the elementary charge. For photoinduced electron

16

II Background and Motivation transfer processes the excitation energy E00 of the photosensitizer has to be considered as well, because it contributes to its redox properties. Electronic Coupling

In addition to the factors introduced above, the reaction rate is also dependent on the probability of the electron transfer, once the system has reached the crossing point from one parabola to the other. This probability is a function of the electronic coupling between donor and acceptor, HAB, more accurately, the overlap between the donor and acceptor orbitals. Evidently, the orbital overlap is maximal for systems with D and A separated only by their van-der-Waals distance; then, HAB = H0. In systems with very strong electronic coupling, differentiation between the reactant and product states is no longer possible (blue parabolas in Figure II.3) and the charge might no longer be assigned to one specific point in the system. It is delocalized as for example in the Creutz-Taube ion.48 

H AB

β = H0 · exp − r DA 2

 (II.10)

rDA in the above Equation II.10 is the donor-acceptor distance. With increasing distance between D and A, the orbital overlap decreases exponentially. For weakly coupled systems (cyan parabolas in Figure II.3) the probability of electron transfer is dependent on the value of EA. Is the excitation energy too low for photoinduced electron transfer to occur, the system will remain in its initial state. β is the attenuation factor for electron tunneling. It is a system-specific parameter, also dependent on the matrix D and A are embedded in. In vacuo, e.g. the probability for electron tunneling is very low.49 In D–A systems linked via covalently attached bridging units, two different mechanistic pathways for long-range electron transfer are possible. If the bridge itself is redox-active the electron or hole can be transferred sequentially via a hopping mechanism. In doing so, intermediates with a temporarily oxidized or reduced bridge are formed. Another possibility is direct tunneling of the charge from the reactant state to the product state. Electron tunneling is usually observed in systems with a high activation barrier.

II.3.3 Energy Transfer Another possibility for deactivation of the excited state besides electron transfer is non-radiative energy transfer from an excited photosensitizer P* to a so-called quencher Q (energy acceptor). For successful electron transfer and thus, prevention of this unwanted side-reaction, the triplet energies of the molecular building blocks

17

II Background and Motivation of the studied complexes have to be adjusted carefully. In energy transfer theory two different mechanisms can be distinguished. An important prerequisite for both mechanisms is a lower-lying energy state of Q, compared to P*. Only then can energy be transferred. Usually the probability for energy transfer can be deduced from the spectral overlap between the energy donor emission and energy acceptor absorption. The higher the spectral overlap, the more likely is energy transfer from one molecular entity to the other. In 1947 T. Förster developed his theory of non-radiative fluorescence resonance energy transfer (FRET).50 FRET is mostly dependent on electronic dipole interactions between the energy donor and acceptor. Its probability decreases with the sixth power of the P*-Q distance. As a result, FRET is the predominant form of energy transfer in systems with P*–Q distances up to 100 Å. LUMO HOMO 3

P*

+

1

Q

1

P

+

3

Q*

Figure II.5 Schematic diagram for electron-exchange triplet energy transfer according to Dexter.

D. L. Dexter formulated his theory of energy transfer on terms of an electron exchange between P* and Q.51 Therefore, sufficient overlap of the orbitals of P* and Q is necessary. Due to this fact, unlike FRET, the electron-exchange energy transfer is more strongly distance-dependent. It usually occurs in systems where P and Q are located within 10 Å of each other, and thus is also termed short-distance energy transfer mechanism. Figure II.5 depicts the mechanism for Dexter energy transfer between an excited P* in its triplet state and the quencher Q in its singlet state in a schematic fashion. This is the most likely energy transfer mechanism to occur in the complexes studied in this thesis. Its probability is dependent on the excitation energies of the present molecular units and will be discussed in the respective sections. The denomination “photosensitizer” for P* can also be explained with the aid of Figure II.5. Electron exchange between 3P* and 1Q leads to the excitation of 1Q to its lowest triplet state. This process is called “sensitization”.

II.4 Accumulative Electron Transfer In a system designed to achieve multiple, consecutive charge separation, new undesired deactivation processes have to be considered that can not occur in molecules de-

18

II Background and Motivation signed for single charge separation. Figure II.6 illustrates the different energy states and electron transfer steps involved in twofold charge separation in an exemplary D–P–A assembly.37,52 The left half depicts the familiar electron transfer pathways resulting in the formation of the first charge-separated state D+PA–(CSS1) after the absorption of one photon (solid, blue arrows). Once CSS1 is formed, its main deactivation occurs via charge recombination (black, dashed arrow), resulting in the DPA triad in its ground state. To obtain a charge separated state in which two electrons and two holes are accumulated on the acceptor and the donor, respectively, the photosensitizer in D+PA– has to be excited again. An important prerequisite hereby is a sufficiently long lifetime of CSS1 to allow for a second excitation. Given successful consecutive absorption of two photons, ideal formation of the desired second charge separated state D2+PA2– (CSS2; solid, cyan arrows) would follow a similar pathway as the formation of CSS1. However, after absorption of a second photon the excited CSS1 D+P*A– can follow deactivation pathways that are not relevant in DP*A. In principle, four major problems are likely to complicate successful accumulative electron transfer.

E

+

D P*A

-

+

+

DPA DP*A

+

2-

-

D PA

2+

D PA +

DP A

hn

-

(CSS2)

hn

+

2-

-

D PA (CSS1)

DPA

Figure II.6 Schematic depiction of possible reaction pathways leading to the accumulation of charge in a DPA triad. Undesired reaction pathways are depicted with dashed arrows.

First, the accumulation of charge, may it be holes or electrons, on one specific reaction site does entail electrostatic problems. It is known that the second oxidation or reduction of one molecular entity is more difficult than the first and requires a higher amount of energy. In a photosensitized assembly, the redox potential of the electron transfer-inducing moiety P* does not change. Therefore, multiple oxidation/reduction of the respective donor/acceptor has to occur in a very narrow po-

19

II Background and Motivation tential range. Only then can P* drive a second charge separation. In principle, the built-up of charge could be compensated, e.g. by coupling it with proton transfer, as seen in PSII and the OEC (Section II.1). A second problem is that the excited photosensitizer P* is both a strong oxidant, as well as reductant. In the D+P*A– state it is surrounded by a reduced, electron-rich acceptor and an oxidized, electron-deficient donor. Thus, P* might be quenched by reverse electron transfer from A– or to D+, as indicated by the dashed, cyan arrows forming D+P–A or DP+A–. Differentiation between the above mentioned reverse electron transfer and back electron transfer is necessary. The latter term is commonly used for direct charge recombination yielding the system in its ground state. Third, the spatial proximity of P* to D+ and A– might also entail undesired quenching of the photosensitizer by energy transfer. The singly reduced acceptors or oxidized donors are mostly radicals or paramagnetic metal centers in their doublet states. Triplet-doublet energy transfer might be less likely in D+P*A– than reverse electron transfer. However, it was observed in other systems and has to be taken into consideration for giving a full evaluation of possible deactivation mechanisms.53 Thus, careful matching of the energies both in DPA, as well as in D+PA– is crucial. Lastly, the intermediate charge-accumulated states, as e.g. D+P+A2– cannot only undergo reverse electron transfer, resulting in CSS1 (dashed, black arrow). An additional deactivation pathway would be direct recombination of electrons and holes. The twofold reduced acceptor A2– can, in principle, transfer an electron to D+, transforming the DPA assembly to an intermediate state of the CSS1 formation. As a result thereof, controlling the electronic coupling between the molecular moieties is of even greater importance than for standard photoinduced electron transfer.

II.5 Artificial Photosynthesis When it comes to artificially creating a system capable of multiple charge separation and the subsequent use of these separated redox equivalents, Nature has provided a working blue-print. It reveals several basic requirements needed for reverse-engineering artificial systems that are as follows. The harvesting of as much sunlight as possible requires suitable chromophores or antennas that display strong absorption in a large part of the solar spectrum. Upon light absorption the energy has to be transferred loss-free to a reaction center which promotes fast charge separation. In order to prevent undesired back-reactions and thus, recombination of the separated charges, spatial separation of the charge accumulation sites is essential. Ideally, the different sites of charge accumulation are redox active and therefore, can

20

II Background and Motivation transfer multiple redox equivalents to catalysts for H2 and O2 evolution, respectively. In addition to the prevention of unwanted back-reactions, the spatial separation of the hydrogen- and oxygen-producing sites does not allow for O2 and H2 to partake in undesired reactions before they can be removed from the system. Lastly, the built-up of charge bears an electrostatic problem. Thus, Nature has coupled the separation of multiple redox equivalents with proton transfer to maintain electric neutrality.54 Up until now several research groups have synthesized molecular systems capable of charge accumulation and a few reviews concerning this topic exist in literature.55–57 Excitation usually occurred by visible light irradiation or with the help of lasers. For most of the systems the use of sacrificial electron donors or acceptors was necessary in order to re-reduce/-oxidize the photosensitizer. However, some of the systems were even reported to produce hydrogen. Hereafter, selected charge accumulating systems will be introduced.

Electron-Accumulating Systems The first electron-accumulating system that did not even rely on a sacrificial electron donor, was presented in 1992 by Wasielewski and co-workers.58 They reported a D– A–D compound comprised of two free-base porphyrin donors (P) and one central perylene bis(dicarboxyimide) acceptor (PBDCI) (Figure II.7). C5H11

C5H11 NH

N

C5H11 N

HN C5H11

O

O

N

N

O

O

NH

N C5H11

N

HN C5H11

Figure II.7 P–PBDCI–P

Excitation of the porphyrins with a low-energy picosecond laser pulse induced the formation of singly charge-separated HP+–PBDCI––PH in pyridine solution. Upon increasing the laser excitation power from one photon per pulse per molecule to 20, the doubly reduced PDBCI unit could be observed after two successive electron transfers. Even the 20-fold increase in the photon-count led to only 20% HP+–PBDCI2––PH+ formation, though quantitative conversion at such high powers was expected. However, charge recombination seemed to occur in the inverted region, since the doubly charge-separated state exhibited a longer lifetime (5 ns) than HP+–PBDCI––PH (5 ps). Six years later, Imahori has communicated a similar charge accumulation system

21

II Background and Motivation with two peripheral zincporphyrin (ZnP) donors and a central tetracyanoanthraquinodimethane (TCAQ) acceptor. Double-pulse excitation beginning with a 532 nm laser pulse, followed by a 233 ps-delayed laser pulse at 555 nm resulted in spectral changes evidential for twofold reduced TCAQ. The obtained second charge-separated state ZnP+–TCAQ2––ZnP+ decayed in only 1.2 ns.59 Other early examples for electron accumulation were reported by Brewer et al. They synthesized molecular heteropolymetallic complexes that showed electron accumulation at the central complex. In a first Ru–Ir–Ru triad (Figure II.8 left), a central iridium complex was attached to two ruthenium complexes via two 2,3-bis(2pyridyl)benzoquinoxaline (dpb) bridging ligands. Visible light irradiation of the triad in MeCN solution lead to twofold reduction of the central iridium complex after absorption of two photons in presence of dimethylaniline as sacrificial electron donor. More accurately, the accumulated electrons are located one on each of the two dpb ligands.Though accumulated at the central iridium complex, the reduction equivalents stay spatially somewhat separated.60 5+

5+ N N N

N Ru N

N N

N

N Ir

Cl

N N

Cl

N

N Ru N

N N N

N Ru

N N

N N

N

N

N Rh

Br

N N

Br

N Ru

N N

N N

Figure II.8 Trinuclear Ru–Ir–Ru (left) and Ru–Rh–Ru (right) complexes.

Brewer et al. also presented structurally similar Ru–Rh–Ru triads (Figure II.8 right) capable of photoinduced electron accumulation on the central rhodium metal complex. The different metal complexes are connected via 2,3-bis(2-pyridyl)pyrazine (dpp) bridging ligands. In presence of dimethylaniline, excitation at 470 nm of a chloro-substituted analog of Ru–Rh–Ru in an acidic MeCN/H2O mixture showed hydrogen production by water oxidation.61–65 Konduri et al. have synthesized two bridged ruthenium dimers, one of which is able to perform photoinduced accumulation of four electrons and four protons on the central bridging unit. Both complexes are comprised of two ruthenium phenanthroline photosensitizers, linked either by a tetraazatetrapyridopentacene (tatpp) or a tetraazatetrapyridoquinone (tatpq) bridge (Figure II.9). Irradiation of a de-oxygenated MeCN solution of the quinone complex with a tungsten light bulb ( λ > 360 nm) resulted in the sequential reduction and protonation of the bridging unit. The generation of the quadruply reduced and protonated form was finished within minutes and proved to be fully reversible upon exposure to

22

II Background and Motivation 4+ N N

N

N N

N

N

N

Ru N

4+

N

N

N

N

N

N

N

N N

N

Ru

Ru

N

O N

N N

[(phen)2Ru-tatpp-Ru(phen)2]

N

N N

N Ru

N

N

N

O

N N

[(phen)2Ru-tatpq-Ru(phen)2]

Figure II.9 Ruthenium Photocatalysts

air.66 A more extensive study of the pyridopentacene containing complex revealed photochemical generation of the doubly reduced and protonated species.67 An excess of NEt3 as a sacrificial electron donor was needed in every case. In contrast to the systems presented above, Fox et al. have reported a purely organic dendritic system comprised of 16 naphthyl groups arranged around one methylviologen unit. Excitation of the compound at 266 nm with strong (> 6 mJ) laser pulses lead to the absorption of about 12 photons by the naphthyl groups, which resulted in twofold reduced methylviologen. Though this system exploits the antenna effect nicely and therefore has no need for an additional sacrificial electron donor, the high excitation density and the very short excitation wavelength limits its usefulness for solar energy conversion.68 Due to the intrinsic difficulties accompanying electron accumulation in purely molecular systems, the focus of current research turned towards the design of hybrid systems. Quite recent examples are phthalocyanine-sensitized single wall carbon nanotubes (SWCNs)69,70 or polyoxometalate (POM)-photosensitizer hybrids.71,72 Continuous visible light irradiation of the latter usually leads to the accumulation of two electrons on the POM moieties that are well suited as charge reservoirs. The POM-containing compounds also exhibited hydrogen production at their headspace, albeit always in presence of sacrificial electron donors such as triethanol- or triethylamine.

Hole-Accumulating Systems Molecular systems capable of hole accumulation on one specific site are quite rare in literature. The ones that can be found are combinations of ruthenium polypyridyl photosensitizers and dinulear manganese clusters, like the ones prepared by Åkermark et al. In presence of [Co(NH3)5Cl]2+ as the sacrificial electron acceptor, accumulation of three holes was observed at the dinuclear Mn-cluster.73,74 Another reported system, comprised of three ruthenium polypyridyl sensitizers bound to a

23

II Background and Motivation [Mn(bpy)3]2+ moiety, can accommodate up to four holes after reorganization to a µoxido-bridged manganese dimer.75,76 Despite their hole accumulation abilities, both systems are stuck in their highly oxidized form due to the lack of oxygen evolution, which prevents them from returning to their initial state.

H3CO

TiO2 O O O

O

N

OCH3

N

OCH3

N N

N

O

N

N

Ru NC

N

NC

H3CO

Figure II.10 TiO2–Ru–OTA

In 2010 Karlsson et al. have presented an oligo-triarylamine (OTA)–Ru-sensitized TiO2 nanoparticle system (Figure II.10). To the extent of the author’s knowledge, it is so far the only example for twofold electron and hole accumulation at two specific sites in the same compound. Furthermore, twofold charge separation was observed entirely without the help of sacrificial agents. After double pulse excitation of the ruthenium photosensitizer, the two electrons donated by the OTA are injected into the nanoparticles. The same result was observed in high-energy single pulse experiments. Though, not a purely molecular system, employment of TiO2 nanoparticles as electron acceptors lead to very fast charge separation and significantly slowed down back electron transfer.52,77 For the determination of the exact whereabouts of both holes and electrons in a system, a molecular, homogeneous approach would be preferable. Furthermore, a more directed electron transfer pathway may be obtained in a linear D–P–A assembly that could then ideally be incorporated directly in an artificial photosynthetic setting as depicted in Figure I.1. An entirely molecular system would allow for a facilitated understanding of the underlying charge transfer, charge accumulation, and fuel generation mechanisms. All reported systems so far have achieved only half reactions where either water oxidation or fuel production is replaced by sacrificial (“dead-end”) reactants. However, the ultimate artificial photosynthetic system has not been realized yet.

24

II Background and Motivation

II.6 Chosen Molecular Moieties II.6.1 Metal Photosensitizers Transition metal complexes, most notably [Ru(bpy)3]2+ or [Re(bpy)(CO)3(py)]+, have been extensively studied and applied as photosensitizers.36,78–81 This can mostly be attributed to their superior chemical and photostability, and their redox properties both in the ground and excited state. Furthermore, their excited states are easily accessible by irradiation, the luminescent states are populated quantitatively and have long lifetimes. By switching from the rhenium photosensitizer to the ruthenium analog, irradiation with visible light is possible which becomes evident from the ground state absorption spectra in Figure II.11.

0.4 20

MLCT

0.2

MC 0

0 400

500

600

wavelength / nm

700

800

extinction coefficient / 103 M-1cm-1

extinction coefficient / 103 M-1cm-1

0.6

40

1.0

25 0.8 20 0.6

LC 15

0.4

10 MLCT

5

0.2

normalized intensity a.u.

0.8

normalized intensity a.u.

60

300

30

1.0

LC

0

0 300

400

500

600

700

800

wavelength / nm

Figure II.11 Absorption (solid trace) and emission (dashed trace) spectra of [Ru(bpy)3](PF6)2 on the left and [Re(bpy)(CO)3(py)](OTf) on the right in MeCN at RT.

[Ru(bpy)3]2+ with Ru2+ as a d6 transition metal, exhibits D3, slightly distorted octahedral symmetry. The absorption spectrum of [Ru(bpy)3](PF6)2 in MeCN at RT (Figure II.11 left) shows a very prominent band at 290 nm which is assigned to a ligand-centered π π* transition. The rather weak absorption band at 350 nm results from a metal-centered transition (MC). The [Ru(bpy)3]2+ 1MLCT band can be seen at 452 nm. The shoulder at 422 nm stems from the slightly distorted geometry of the metal center.36,82 The rhenium metal center in [Re(bpy)(CO)3(py)]+ also exhibits approximately octahedral coordination symmetry. The absorption spectrum of this complex shows bands of the same general origin as the ruthenium polypyridyl spectrum. The high-energy bands from 250 nm to the sharp band at 320 nm are assigned to π π* transitions centered on the bpy ligand. The 1MLCT band for the dπ(Re) π*(bpy) transition is located at higher energies compared to [Ru(bpy)3]2+. It is partially overlapped by the ligand-centered transitions, but clearly visible from 330 nm to longer wavelengths and tailing into the visible region.80

25

II Background and Motivation After excitation at 450 nm for the ruthenium complex and 350 nm for the rhenium sensitizer, both metal complexes show intense, broad, structureless emission spectra at RT in MeCN. The emission maximum for [Ru(bpy)3]2+ in MeCN is located at 603 nm, whereas the maximum rhenium emission is further into the blue at 553 nm in the same solvent. Light-induced excited state formation and electrochemical characteristics of the metal complexes in their ground and excited states can be summarized in a Latimer diagram (Figure II.12). Excitation of the metal photosensitizers usually occurs into their 1 MLCT state and ISC to the actual luminescent 3MLCT state occurs quantitatively. The excited state lifetime is 0.8 µs for [Ru(bpy)3]2+ in de-oxygenated MeCN solution with an emission quantum yield of 0.06.83 The emissive state of [Re(bpy)(CO)3(py)]+ has a lifetime of 0.6 µs in de-oxygenated DCM and φem = 0.16.84 1

Ru(bpy-)(bpy)22+ F=1

3

Ru(bpy-)(bpy)22+ 2.12 V

hn

0.84 V

Ru(bpy)3+

- 1.28 V

1

hn*

Ru(bpy)32+

- 0.86 V 1.26 V

Ru(bpy)33+

Re(bpy-)(CO)3(py)+ F=1

3

Re(bpy-)(CO)3(py)+ 2.82 V hn

1.64 V

Re(bpy)(CO)3(py)0

- 1.18 V

hn*

Re(bpy)(CO)3(py)+

- 1.07 V 1.75 V

Re(bpy)(CO)3(py)2+

Figure II.12 Latimer diagrams for the metal photosensitizers.

After excitation into the 1MLCT band, the excited metal polypyridyl complexes are both stronger reductants, as well as oxidants than in their ground states. This can be explained by their higher energy content, compared to the ground state. The connection between ground state redox potentials and the zero-zero excitation energy (E00) is visualized in the Latimer diagram. Redox potentials are given in V vs. NHE in H2O for [Ru(bpy)3]2+ and in V vs. SSCE in MeCN for [Re(bpy)(CO)3(py)]+.54,84 An estimation of the excited state redox potentials is possible with the help of the following Equations II.11. Here, the ground state potentials and the zero-zero excitation energy have to be taken into account.85

26

II Background and Motivation

E( M+ /M∗ ) = E( M+ /M) − E00 E( M∗ /M− ) = E( M/M− ) + E00

(II.11)

II.6.2 Anthraquinone Acceptor Nature has successfully used quinones as the final electron acceptors in PSII. Following this example, quinone derivatives, such as anthraquinone, have already been employed as electron acceptors in photoinduced charge transfer studies.86–89 Anthraquinone is able to accommodate two electrons and its electron accepting behavior has been studied extensively.90–93 Electrochemical studies have shown two reversible one-electron reduction waves for the first (−1.20 V vs. Fc+/0) and second reduction (−1.8 V vs. Fc+/0) in DMF. The second reduction occurs at a potential that is too low to be reached by the reduced/oxidized rhenium or ruthenium photosensitizer. However, addition of Brønsted or Lewis acids or strongly hydrogen bond-donating solvents (HFIP) leads to merging of the two separate reductions, forming a new two-electron reduction wave at a more positive potential. Protonation of AQ– is usually followed by instant disproportionation of AQH to twofold reduced and protonated AQH2 and AQ, because AQH is more readily reduced than AQ. Together with the positive shift of the second reduction potential, the reversibility of the two-electron reduction wave around −1.20 V vs. Fc+/0 is crucial for achievement of twofold charge separation in the rhenium and ruthenium containing triads reported below.

II.6.3 Oligo-triarylamine Donor Oligo-triarylamine (OTA) was chosen as the electron donor, due to its formidable electron donating abilities. It can release two electrons within a very narrow potential range of about 0.3 V. Since it displays very low potentials of about −0.05 and 0.25 V vs. Fc+/0 for the first two oxidations, it is a much stronger electron donor than its ordinary triarylamine analog. The first oxidation of OTA occurs at the central N atom, the most electron-rich position in the donor, and the charge of the resulting radical cation can be stabilized due to the extended ring system.94 Due to its superior electron-donating properties, OTA has already been employed successfully in donor–metal photosensitizer–acceptor assemblies as well as in organic hole transport materials.52,95,96

27

III. Towards Photoinduced Charge Accumulation in d6 Metal Complexes

Electron transfer plays a pivotal role in many biological and physiological processes and the elucidation of its mechanism and dependency on various parameters, as well as the creation of artificial electron transfer systems has received a lot of attention in the last decades. More recently the focus of electron transfer studies turned towards the accumulation of charges. With solar fuel generation in mind, more than one, even multiple redox equivalents are required for successful H2 production or CO2 reduction. Several purely molecular systems capable of either hole or electron accumulation with the help of sacrificial agents have been reported. However, none so far that can accumulate both holes and electrons at two specific sites in a completely molecular setup. Therefore, the primary objective of the research in this thesis was to bridge this gap and to synthesize model systems, built as simple as possible, to study possible photoaccumulation of redox equivalents on specific sites within the same purely molecular (nanoparticle-free) compound. A major advantage of such systems will be their independence of the high-energy sacrificial agents and thus, their self-sufficiency. In principle, the initial target molecule, comprised of an OTA donor and two rhenium photosensitizers should allow us to accumulate two holes on OTA with each of the rhenium complexes being reduced once. In Section III.2, a slightly more extended, yet every bit as simple a system (AQ–Re–OTA) will be studied for the same goal. The first part of this chapter will focus on the investigation of the rhenium complexes, what difficulties we encountered, what we learned, and what changes we made in order to achieve our goal of accumulating redox equivalents. The second part will give an overview of the results gained from an AQ–Ru–OTA triad (III.4) and the corresponding pentad AQ–Ru–OTA–Ru–AQ (III.7) and how they differ from the ones obtained from the rhenium triads. In all following sections synthesis, spectroscopic measurements and efforts on our side to achieve photo-accumulative electron transfer will be discussed in detail.

29

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.1 Triads with Two Rhenium Photosensitizers The triads studied in this section are each comprised of two rhenium chromophores covalently linked to either an oligo-triarylamine (OTA) or a triarylamine (TAA) donor. The latter triad, in which hole accumulation is impossible, was used as a reference complex.

III.1.1 Synthesis of the OTA Donor The oligo-triarylamine (OTA) which was used as an electron donor throughout this thesis was synthesized by adapting literature protocols.95,97 The reaction sequence for its preparation is shown in Scheme III.1. Of this three step synthesis, the first step was performed exactly as Hirao et al. described. Commercial bis(4-bromophenyl)amine, DMAP and Boc2O were refluxed in THF which yielded the Boc-protected amine (OTA–pre) quantitatively after the workup procedure. This step was necessary to prevent unwanted side reactions in the subsequent Buchwald–Hartwig coupling with dianisylamine in order to obtain the Boc-protected OTA (OTA–Boc). The synthetic procedure of Hirao et al. was changed insofar that Pd(dba)2 and (HPtBu3)BF4 were used as catalysts. In this way, the reaction time could be reduced from 5 d to 19 h and OTA–Boc was obtained in a superior yield of 84%. Deprotection occured in an acetone solution with an excess of TFA at RT as described by Bozic-Weber et al. and gave the desired OTA in 99% yield. The donor was obtained as a beige solid with an overall yield of 83%. In order to obtain a well resolved 1H NMR spectrum of OTA and thus enabling a clear assignment of signals, the addition of a drop of TFA to the NMR sample was necessary. This resulted in the solution turning green, most likely due to oxidation of some of the amine.98 However, OTA could be recovered by evaporation of the solvent and filtration of the NMR sample over basic alumina with EtOAc as eluent. OCH3 Br Br

Br N H

a) 100 %

OCH3 OCH3

Br b)

N O

O

84 %

N N O

OTA-pre

c)

N

H3CO

OCH3 O

OTA-Boc

OCH3

N

99 % H3CO

N N H

OCH3

OTA

Scheme III.1 Synthesis of the OTA donor: a) Boc2O, DMAP, toluene, reflux, 3 h; b) dianisylamine, NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, reflux, 19 h; c) TFA, acetone, RT, over night.

30

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes In addition to the characterization by 1H NMR spectroscopy a crystal structure (Figure III.1) of the donor was obtained thanks to single crystal formation by slow evaporation of a saturated acetone solution. The structure was deposited at the Cambridge Crystallographic Data Centre (CCDC) under the reference number 1011917. Directly observable from the crystal structure is the propeller-like array of the aryl rings around the nitrogen atoms. This is observed for many arylamines and contributes greatly to their stability when oxidized, and their rather low basicity.99,100

~9 Å

~21.6 Å

Figure III.1

~5 Å

Molecular structure of the OTA donor. ADPs: 50% thermal ellipsoid probability. Carbon atoms: black; hydrogen atom: grey; nitrogen atoms: blue; oxygen atoms: red. (Symmetry operation used to generate equivalent atoms: 0.5-x, 0.5-y.)

OTA crystallizes in the orthorhombic space group Pccn. The distance between C20 and C20’ is 21.6 Å and the through-space distance measured from C13 to the hydrogen atom as indicated in the side view of the donor in Figure III.1 is ≈ 9 Å. Considering the depth of 5 Å, OTA shows ellipsoid geometry.

III.1.2 Synthesis of Precursors and the Reference Molecules To aid the complete understanding of the data obtained from the target complexes, reference compounds comprised of the individual molecular building blocks of Re2– TAA and Re2–OTA were synthesized. The known Re–ref complex, [Re(bpy)(CO)3(py)]+, could be obtained in three synthetic steps as shown in the upper third of Scheme III.2 with an overall yield of 74%. The procedures followed hereby were adapted from literature.80,101,102 Re(CO)5Cl and 2,2’bipyridine (bpy) were stirred in refluxing toluene for 1 h and after cooling the reaction mixture in the fridge the yellow product was filtered and washed with Et2O. Substitution of the chloro-ligand in the [Re(bpy)(CO)3Cl] precursor by the weakly coordinating triflate ligand was performed to facilitate the following complexations with the amino-decorated ligands. For the exchange, [Re(bpy)(CO)3Cl] was suspended in DCM and trifluoromethanesulfonic acid was added very slowly. Precipitation of the

31

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes product from the clear mixture was induced by drop wise addition of Et2O. Finally, the [Re(bpy)(CO)3(OTf)] precursor was refluxed in MeOH, possibly forming the intermediate complex [Re(bpy)(CO)3(MeOH)]OTf which then reacts with an excess of pyridine to give Re–ref.

+ OTf ReCO5Cl +

N

a) N

N

97 %

N

OC Re Cl OC

b) 90 %

CO

N

N OC

N

c)

OC Re OTf

85 %

N

OC Re N

CO

OC

CO

Re-ref H3CO H3CO

OCH3 N H

OCH3

d)

N

84 %

TAA-ref

OCH3

OCH3

N

N

OCH3 e)

H3CO

N H

70 % OCH3

OCH3

N H3CO

N N

OCH3

OTA-ref

Scheme III.2 Synthesis of the reference compounds: a) toluene, reflux, 1 h; b) trifluoromethanesulfonic acid, DCM, RT, 1 h; c) pyridine, MeOH, 85 °C, 24 h; d) bromotoluene, NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, reflux, 23 h; e) bromobenzene, KOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, reflux, over night.

The secondary amines TAA and OTA were reacted with p-bromotoluene and bromobenzene, respectively, in Buchwald–Hartwig-type couplings to yield the tertiary arylamines. The Pd-catalyzed N-C coupling of dianisylamine with bromotoluene yielded the desired TAA–ref in 84% yield. The initially beige oil crystallized within a few weeks. For the Pd-catalyzed synthesis of OTA–ref a procedure of Hirao et al. was adapted.94 During the work-up and purification steps absence of acidic substances proved to be crucial. Recrystallization from MeOH, washing and thorough drying afforded pure OTA–ref as a beige solid in 70% yield. The resulting TAA–ref and OTA–ref were used to help predict the optical spectroscopic and electrochemical behavior of the amines in the complexes. Furthermore, OTA–ref was needed to determine the triplet energy of OTA due to the lack of literature values.

32

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.1.3 Synthesis of the Studied Triads Scheme III.3 outlines the synthetic pathway pursued to obtain the desired Re2–OTA complex and the corresponding complex Re2–TAA which was used as a reference. H3CO H3CO

N

OCH3 N

b)

c)

83 %

50 % N

OC OC

N

CO Re N

N

N

N

N

py-TAA-py

Br Br

2+ 2NO3-

OCH3

CO Re N

CO CO

Re2-TAA

Br a) Br 36 %

N

N

py-Br-py OCH3

71 %

H3CO

N N

e)

N

N

OCH3

H3CO

37 % OC

N

OCH3 2+ 2NO3

OCH3

N

d)

OCH3

N

py-OTA-py

OC

N CO Re N

OCH3

N

N

N

N

CO Re N

CO CO

Re2-OTA

Scheme III.3 Synthesis of the studied complexes Re2–OTA and Re2–TAA: a) pyridine4-boronic acid hydrate, Na2CO3, Pd(PPh3)4, THF/H2O, reflux, 24 h; b) dianisylamine, NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, 100 °C, 22.5 h; c) [Re(bpy)(CO)3OTf], MeOH, 85 °C, over night; d) OTA, NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, reflux, 19 h; e) [Re(bpy)(CO)3OTf], MeOH/CHCl3, reflux, 1.5 d.

The first step towards the complexes consisted of the synthesis of a central molecule (py–Br–py) used for both, the Re2–TAA and Re2–OTA complex. It is comprised of a benzene ring bearing two pyridines as coordination sites for the rhenium precursor ([Re(bpy)(CO)3OTf]) and a bromine substituent which allowed for the coupling of the respective amine. 1,3,5-tribromobenzene was reacted with 3 eq of 4-pyridine boronic acid in presence of Na2CO3 as base and catalytic amounts of Pd(PPh3)4 in a biphasic solvent mixture of THF and H2O, as is common for Suzuki-type couplings. The use of 3 eq of the pyridine reactant seemed to be crucial for the conversion of 1,3,5-tribromobenzene to the desired compound in an acceptable yield. Still, py–Br– py could only be isolated with a yield as low as 36%. Purification of the product was achieved by column chromatography on silica with DCM + 1% NEt3 as eluent. The 1H NMR spectrum in CDCl3 showed three signals (s, 5.66 ppm; q, 3.64 ppm; t,

33

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 1.48 ppm) in addition to the product signals. Those signals are most likely due to a decomposition product of NEt3 in the weakly acidic DCM. This was concluded, because the signals were absent after column chromatography in DCM with MeOH. However, only insufficient purity could be achieved with the latter eluent mixture. The decomposition product was always present throughout the next synthetic steps, but could be successfully removed in the purification steps of the complexes. Due to a lack of participation in the following synthetic steps and its possible removal during complex preparation and purification the decomposition product was tolerated. Coupling of OTA with py–Br–py was performed under the same conditions used for the synthesis of OTA–Boc as shown above and produced py–OTA–py in 71% yield. The combination of Pd(dba)2 and (HPtBu3)BF4 as catalysts with either NaOtBu or KOtBu as base in toluene consistently provided the best yields in the performed N-C couplings and was therefore used as the standard reaction condition. Complexation of the py–OTA–py ligand with [Re(bpy)(CO)3OTf] in a MeOH/CHCl3 solvent mixture produced Re2–OTA, as well as the singly complexed by-product. The desired product was isolated by column chromatography on silica by first eluting unreacted ligand with pure acetone, then the complex was eluted with acetone/H2O/sat. KNO3 200:9:1 (v:v:v). After evaporation of the acetone the complex was extracted from the aqueous phase with DCM, dried and further recrystallized from hexane. This gave Re2–OTA as the nitrate salt in form of an orange solid with an overall yield of 9%. Because the synthesis of the reference complex Re2–TAA was performed in analogy to Re2–OTA it will not be discussed in detail. The upper half of Scheme III.3 shows the reaction sequence following which Re2–TAA was synthesized in three steps starting from the commercially available 1,3,5-tribromobenzene. The complex was obtained as a nitrate salt with an overall yield of 15%.

III.1.4 Characterization by IR and Optical Absorption Spectroscopy For spectroscopic characterization, IR spectra and UV-Vis absorption spectra of all the complexes were measured. The IR spectra were directly obtained from the solids, whereas the absorption spectra were measured on 10−5 M solutions of the complexes in MeCN. The IR spectra (Figure III.2) show two prominent bands corresponding to vibrations of the CO-ligands on the rhenium photosensitizers. The band at 2030 cm–1 is assigned to the totally symmetric in-phase vibration and the one at 1900 cm–1 to the out-of-phase totally symmetric vibration superimposed by the asymmetric stretching of the equatorial CO-ligands.103 The fingerprint regions, especially of Re2–TAA and Re2–OTA, show identical bands, because both complexes are comprised of the same functional

34

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes groups (N-heterocycles, methoxy groups, and tertiary arylamines).

2200 2000 1800 1600 1400 1200 1000

wavenumbers / cm

Figure III.2

800

600

400

-1

IR spectra of the investigated complexes Re–ref (black), Re2–TAA (blue), and Re2–OTA (cyan).

extinction coefficient / 103 M-1cm-1

In Figure III.3 the optical absorption spectra of the investigated complexes are depicted. The spectrum of the Re–ref complex is as expected and coincides with that reported in the literature.80 The high-energy bands at 260-320 nm are assigned to ligand-centered π π* transitions and the ones at 340-390 nm to a dπ(Re) π*(bpy) (1MLCT) transition. For the investigated complexes Re2–OTA and Re2–TAA the same assignment is applicable. However, the 1MLCT band at 350 nm is less clearly observable in the TAA complex and completely superimposed by OTA absorption in Re2–OTA. UV-Vis spectra recorded on TAA–ref and the OTA–ref show absorption of the amines at 355 nm with extinction coefficients of 1700 M–1cm–1 and 24000 M–1cm–1, respectively. As a result, compared to Re–ref with  = 4700 M–1cm–1, Re2–TAA and Re2–OTA exhibit very large extinction coefficients at 355 nm ( = 16300 M–1cm–1 and  = 36000 M–1cm–1), not unusual for systems with an extended π-conjugation. 80 60 40 20 0 300

400

500

600

700

800

wavelength / nm

Figure III.3

UV-Vis absorption spectra of the investigated complexes Re–ref (black), Re2–TAA (blue), and Re2–OTA (cyan).

Considering the information gathered from the absorption data it is to be expected that, by exciting at 355 nm with a Nd:YAG laser, the amines will also be excited to a significant extent. However, instant relaxation of their excited state (π*) and energy transfer to the excited state (3MLCT) of the photosensitizer will result in the systems

35

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes in their expected photoexcited state. Therefore, this circumstance should have little effect on the formation of possible photoproducts.

III.1.5 Determination of the Redox Potentials and Establishment of the Energy Level Diagrams The redox potentials of the individual molecular moieties in the complexes were determined by means of cyclic voltammetry. All cyclic voltammograms (Figure III.4) were recorded in de-oxygenated MeCN solutions at complex concentrations of about 1 mM and are referenced versus the ferrocenium/ferrocene redox couple Fc+/0. 0.1 M TBAF6 was used as supporting electrolyte, and for all measurements the potential scanning rate was 0.1 V/s. [Re(bpy)(CO)3(py)]+ was also measured as an unsubstituted reference complex, in order to evaluate changes in the redox potentials due to attachment of the amine moieties. Measuring reduction and oxidation in two separate sweeps proved to provide higher quality results for Re2–TAA and Re–ref.

Figure III.4 Cyclic voltammograms of Re2–OTA (cyan), Re2–TAA (blue), and Re–ref (black) in de-oxygenated MeCN solutions referenced to Fc+/0.

The direct use of decamethylferrocene (Me10Fc) as an internal standard for the measurement of Re2–OTA produced data which was not evaluable due to interactions with the complex. More precisely, the redox chemistry of Me10Fc became irreversible with the appearance of new waves not belonging to the complex itself and the oxidation waves for OTA increased with every redox cycle and seemed to merge, thus prohibiting a clear determination of the proper values. To assure calibration of the potentials the following procedure was applied. First, only the supporting electrolyte dissolved in MeCN was measured to establish a baseline, then Re2–OTA was added, and the cyclic voltammograms were recorded. After that, Me10Fc was added and the sample was measured with the same settings again. In the data evaluation, the wave corresponding to Me10Fc was shifted to reference the voltammogram to the potential of the Fc+/0 couple. The voltammograms without the internal standard were shifted

36

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes by the exact same value. Table III.1 summarizes the redox potentials as extracted from the cyclic voltammograms in Figure III.4. Table III.1 Redox potentials in MeCN. All values are given in V vs. Fc+/0.

complex

Re2+/+ /V

bpy0/– /V

Re–ref Re2–TAA Re2–OTA

1.37 1.23 0.88

−1.77 −1.77 −1.88

amine+/0 amine2+/+ amine3+/2+ /V /V /V 0.46 −0.07

0.25

0.65

The literature potentials for Re–ref given in V vs. SCE in MeCN were converted to V vs. Fc+/0 according to Pavlishchuk et al.104 The reduction with Ered = −1.77 V is ligand-centered, whereas the oxidation is metal-centered (Eox = 1.37 V).80 The measured values are in good agreement with literature. Re2–TAA shows one quasireversible oxidation at 0.46 V for the oxidation of TAA to TAA+. Re2–OTA on the other hand, shows three oxidation waves at −0.07 V, 0.25 V, and 0.65 V with the first two reversible. These values coincide well with literature ones obtained from molecules comprised of oligo-triarylamine moieties.77,96 The very low oxidation potentials at around 0 V indicate that, in principle, OTA can easily be oxidized by the excited rhenium photosensitizer. Therefore, OTA is a suitable electron donor and, in the present complex Re2–OTA, can readily accumulate two holes transferring one electron to either one of the rhenium complexes. Interestingly, the CV data shows that the rhenium oxidation potential is shifted to less positive values starting at 1.37 V for the unsubstituted Re–ref complex, to 1.23 V for the TAA substituted complex and down to 0.88 V for Re2–OTA. This behavior is due to the increasing electron-donating ability of the attached amines. The electron donating resonance effect (+M effect) of the methoxy group, if attached in para-position, surpasses its electron withdrawing effect (−I effect). 4-[bis(4-methoxyphenyl)amino]phenyls in general possess a high π-donor strength, a measure for their ability to act as intramolecular electron donors towards conjugated π-systems.105 Hence, the attached amines push negative charge towards the rhenium metal center facilitating its oxidation. This effect, already influencing the oxidation potential of rhenium in Re2–TAA, will most certainly be stronger in Re2–OTA. Due to the more extended arylamine system with three nitrogen atoms each bearing a lone pair and the four attached methoxy groups, OTA is even more electron-rich than TAA. The ancillary bpy-ligands are connected to the rhenium center via coordinative bonds and are not covalently linked to the amines. Furthermore, the distance between bpy and amine is larger which also contributes to their reduction potential not being influenced in such a way. As a result,

37

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes all complexes show similar values for the bpy-centered reduction around −1.8 V. Establishment of the Energy Level Diagrams

The values for the energies of the different redox states were estimated using the following Equation (III.1).106 Errors of this estimation typically lie in the range of 0.1 eV. 0 ∆GET = e ( Eox − Ered ) − E00 −

with Eox Ered E00 e e0 es R DA

e2 4π · e0 · es · R DA

(III.1)

oxidation potential reduction potential 3 MLCT energy elementary charge vacuum permittivity dielectric constant of MeCN (35.94) donor-acceptor distance (≈ 12Å).

The D–A distance of 12 Å was estimated from the distance between the ancillary bpy ligand on the rhenium photosensitizer over the central benzene ring to the first nitrogen atom on the amine donor. The resulting energy level diagrams for both, Re2–TAA and Re2–OTA are depicted in Scheme III.4.

Scheme III.4 Energy level diagrams for Re2–TAA (left) and Re2–OTA (right).

As can be seen, the electron transfer step from TAA to the rhenium photosensitizer

38

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes in Re2–TAA is exergonic by 0.60 eV and therefore very likely to occur, at least from a purely thermodynamic point of view. For Re2–OTA, the transfer of the first electron from OTA to rhenium is exergonic by 1.03 eV. Likewise, the second electron transfer to form OTA2+ and a second reduced rhenium sensitizer is exergonic by 0.71 eV, but only after absorption of a second photon. The thermodynamic data therefore suggests formation of the second charge-separated state in Re2–OTA is indeed possible. This is mainly due to the very low first two oxidation potentials of OTA. Nevertheless, formation of OTA2+ does still require a two-photon process which is not only dependent on thermodynamics.

III.1.6 Steady State Luminescence Spectroscopy The luminescence spectra of the investigated complexes show a broad emission band as displayed in Figure III.5. Rhenium polypyridyl complexes are known to give unstructured emission bands especially at RT which are attributed to forbidden electronic transitions from the 3MLCT excited state to the ground state.80,107,108

relative intensity a.u.

1.0 0.8 0.6 0.4 0.2 0 400

500

600

700

wavelength / nm

Figure III.5

Luminescence spectra of the investigated complexes (Re2–TAA in blue and Re2–OTA in cyan) measured on de-oxygenated 10−5 M MeCN solutions after excitation at 355 nm. All spectra were corrected for differences in extinction at 355 nm and normalized to the emission of Re–ref (black).

Re–ref shows an emission maximum at 553 nm, whereas the maxima of the TAA and the OTA complex are blue-shifted by 12 nm. Re2–TAA and Re2–OTA show significant quenching of the rhenium luminescence. For Re2–TAA it is quenched by 95% and for Re2–OTA by 97.5% which is consistent with relatively strong electronic coupling between rhenium and the amines, due to their spatial proximity. This strong 3MLCT quenching is indicative of photochemistry in the investigated complexes. However, apart from the desired electron transfer from the amines to the rhenium chromophores, energy transfer might be an unwanted non-radiative deactivation pathway. The possibility of this will be discussed in the following subsection.

39

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.1.7 Possibility of Triplet-Triplet Energy Transfer In order to assess the efficiency of energy transfer from the excited rhenium photosensitizers to the amines which may prevent charge separation by fast deactivation of the excited 3MLCT state, one must know the energies of the lowest-lying triplet states of the involved molecular entities. On account of lacking information on the triplet state energy of OTA it was experimentally determined by recording the luminescence spectrum of OTA–ref in a frozen glass (MeTHF, 10−5 M) at 77 K (Figure III.6 b). The triplet energy of Re–ref was measured accordingly, albeit in a EtOH/MeOH 4:1 glass matrix (Figure III.6 a). ET (in eV) then corresponds to the luminescence origin at this temperature. 2.5

emission intensity / 106 a.u.

emission intensity / 106 a.u.

a) 4.0 3.0 2.0 1.0

b)

2.0 1.5 1.0 0.5 0

0 400

450

500

550

600

400

650

450

500

550

600

650

wavelength / nm

wavelength / nm

Figure III.6 Steady-state luminescence spectra at of Re–ref (a), in an EtOH/MeOH 4:1) and OTA–ref (b), in MeTHF) at 77 K.

Excitation occurred at the respective absorption maxima, i.e. 340 nm for OTA–ref and 355 nm for Re–ref. The ET values obtained, summarized in Table III.2, are 3.32 eV for OTA an 2.82 eV for Re–ref. The latter is comparable to the value that can be extracted from the emission spectrum at 77 K shown by Sacksteder et al.80 Furthermore, a value of 3.04 eV for the lowest triplet excited state of TAA can be found in literature.109 Table III.2 Luminescence origin and triplet energies.

Re–ref OTA–ref

origin / nm

ET / eV

440 373

2.82 3.32

On the basis of the above determined values, it becomes clear that triplet energy transfer from the excited rhenium chromophore to TAA is endergonic by 0.22 eV and to OTA even by 0.5 eV. Consequently, quenching of the excited state most likely occurs by electron transfer and not energy transfer. Direct evidence for the importance of electron transfer will be provided in III.1.9.

40

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.1.8 Time-Resolved Luminescence Spectroscopy

normalized optical density

To gain further insight into the photochemical processes occurring in the Re2–OTA and Re2–TAA triads, time-resolved emission measurements were performed to determine at what rate luminescence was quenched. In de-oxygenated 10−5 M MeCN solutions of the triads immediate quenching of the luminescence after laser excitation (pulse duration ≈10 ns) was observed. This corroborates the results from the steadystate measurements. Exponential fitting of the decay gives a luminescence lifetime of 239 ns for Re–ref. The rhenium triads show a biexponential decay indicating the existence of two different luminescent species. The data for Re2–TAA suggests one species whose luminescence decays with τ = 21 ns representing the main product with a contribution of 76% and another product (τ = 332 ns) contributing only to 24%. In the case of Re2–OTA, the values present themselves as follows: short-lived luminescent species with τ = 27 ns (64%) and long-lived luminescent species with τ = 423 ns (36%). 1

0.1

0.01

0.001 0

200

400

600

800

1000

1200

time / ns

Figure III.7

Luminescence decays of Re–ref (black), Re2–TAA (blue), and Re2–OTA (cyan) in de-oxygenated MeCN solutions (10−5 M) after excitation at 335 nm.

These findings might be caused by the presence of two rhenium complexes in these triads. By excitation with a laser pulse at 335 nm both rhenium chromophores in one complex could be excited simultaneously. Following the luminescence quenching of one by electron transfer from the amine to the metal center, the other chromophore might still be emissive. The longer lifetimes in the triads need not match the luminescence lifetime of the free Re–ref exactly, due to a changed electronic and possibly even structural environment following the charge separation. An interpretation such as this seems much more valid for Re2–TAA, because here the amine can only be oxidized once and hence quench solely one excited metal center. Re2–OTA, however, was designed to quench both rhenium photosensitizers by two successive oxidations of OTA. In the present measurement the photon flux density might have been too low to accomplish the required two-photon process, leading to a mixture of two

41

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes different photoproducts. Quadratic excitation power dependency is a well known characteristic of two-photon processes.110,111 An alternative interpretation would be that the biexponential luminescence decays in the rhenium triads are attributable to minor, yet highly luminescent impurities. Equation III.2 was applied to the luminescence lifetimes of the short-living species of Re2–TAA and Re2–OTA in order to obtain the quenching constants for both complexes in MeCN solution. k Q = τobs −1 − τre f −1

(III.2)

kQ is the quenching constant calculated from the luminescence lifetimes of Re2–TAA –1 and Re2–OTA, respectively (τobs ), and the luminescence lifetime of the unsubstituted –1 reference complex [Re(bpy)(CO)3(py)]+ (τref ). –1 7 For Re2–TAA a kQ-value of 4.34·10 s was calculated, as compared to kQ = 3.29·107 s–1 for Re2–OTA.

III.1.9 Identification of the Photoproducts With both, quenching of the steady state rhenium emission and shortening of the excited-state lifetime being observed, it seems certain that photochemistry in form of electron transfer is indeed occurring in the Re2-triads. The occurrence of energy transfer as a possible deactivation pathway of the excited state was found to be unlikely by estimating the triplet energies of the different molecular moieties (Subsection III.1.7). Based on the above mentioned data for Re2–TAA, upon excitation of a de-oxygenated MeCN solution electron transfer from the TAA moiety to one rhenium photosensitizer is expected. In order to get an idea of what to expect for TAA+ in transient absorption spectra, chemical oxidation studies have been made by titrating an MeCN solution of the complex with Cu(ClO4)2. Figure III.8 shows the UV-Vis-absorption data of the titration leading to Re2–TAA+ (top), as well as the transient absorption spectrum after excitation at 355 nm (bottom). The transient spectrum of Re2–TAA, recorded in a 200 ns time window directly after the excitation pulse shows two bands at 771 nm and 363 nm attributable to the formation of the TAA radical cation. These are typical features observed for oxidized TAA.112,113 Reductive flash-quench experiments of rhenium polypyridyl complexes with amine quenchers, such as DPA or TEOA, indicate that the positive absorption

42

mDOD

absorbance

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.6 0.5 0.4 0.3 0.2 0.1 0 8 6 4 2 0 -2

a)

b)

300

400

500

600

700

800

wavelength / nm

Figure III.8

Top: UV-Vis absorption spectrum of a 10−5 M MeCN solution of Re2–TAA (black) and its stepwise oxidation to Re2–TAA+ (blue) upon addition of increasing amounts of Cu(ClO4)2; bottom: transient absorption spectrum after 355 nm excitation of a 35 µM de-oxygenated MeCN solution of Re2– TAA.

normalized optical density

around 450-500 nm is due to the reduced rhenium photosensitizer.114 A bleach at 320 nm is also expected for the reduced rhenium complex.108,115 However, this is not observed here, presumably because of the rather high optical density of the complex solution in this spectral range. Given the direct evidence for the formation of a charge-separated state with a hole on the TAA and a negative charge on the rhenium photosensitizer, or more exactly on the ancillary bipyridine ligand, we expect both radical species to decay with the same time constant. The decay of the photoproduct to its ground state as detected at different wavelengths is displayed in Figure III.9, and it can be seen that this expectation is indeed met. The charge-separated photoproduct (Re+)(Re0)–TAA+ is generated within the 10 ns duration of the laser pulse and decays within 39 ns. Though not a very longliving species, this short lifetime could be expected for rather closely connected donor and acceptor moieties, as is the case in the present molecule. 1 360 nm 490 nm 770 nm

0.1

0.01 0

50

100

150

200

time / ns

Figure III.9

Decay of transient absorption signals at 360 nm (blue), 490 nm (cyan) and 770 nm (orange).

In contrast to Re2–TAA, twofold charge separation to yield (Re0)2–OTA2+ with two holes on the OTA donor and one electron at each of the two rhenium acceptors could be anticipated. To test the possibility of achieving the necessary two-photon

43

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

absorbance

absorption an ordinary pump-probe technique was applied. The laser pulse duration was ≈10 ns. From the experiments on Re2–TAA it is known that the photoproduct is formed within the 10 ns laser pulse since no rising signal could be detected. With OTA having such low oxidation potentials that it can be oxidized twice until the potential to oxidize TAA to TAA+ is reached, we hoped to gain access to the twofold charge-separated state of Re2–OTA. The lifetime of the first charge-separated state, (Re+)(Re0)–OTA+, is estimated to be in the same order of magnitude as for Re2–TAA which is long enough to excite one molecule two times within the duration of the laser pulse. In order to understand the changes in the absorption of Re2–OTA upon undergoing photoinduced electron transfer, it was chemically oxidized by titration with Cu(ClO4)2. Figure III.10a depicts the data collected by stepwise chemical oxidation and excitation with a 10 ns laser pulse at 355 nm (Figure III.10b). On the basis of the data obtained hereby it becomes clear that OTA+ exhibits positive absorption throughout the whole spectral range from 400 to 800 nm with local maxima at 427 nm and 765 nm and a very broad absorption band in the NIR region with a maximum at around 1350 nm. When Re2–OTA+ is further oxidized to Re2–OTA2+, the absorption at 427 nm decreases to approximately one third of the initial OTA+ absorption with a concomitant increase between 500 and 650 nm resulting in a new absorption maximum at 600 nm. Furthermore, the absorption at 765 nm is also still increasing, albeit only slightly. However, the most distinctive difference between Re2–OTA+ and its twofold oxidized form can be seen in the NIR region. There, a new and very prominent band at 1138 nm arises which can undoubtedly be assigned to Re2–OTA2+.

0.6

0.2 0 4

mDOD

a)

0.4

b)

2 0 -2 400

600

800

1000

1200

1400

1600

wavelength / nm

Figure III.10 Top: stepwise chemical oxidation of a 10−5 M MeCN solution of Re2–OTA (black) to Re2–OTA+ (blue) and Re2–OTA2+ (cyan); bottom: transient absorption spectrum of a de-oxygenated 10−5 M MeCN solution of Re2–OTA directly after 355 nm excitation in a 50 ns time window.

A definite distinction between the two oxidation states of OTA in the spectral range from 400 to 800 nm may prove difficult due to their similar absorbance in that area.

44

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

normalized optical density

Therefore, in addition to the transient absorption data acquired by a PMT between 300 and 800 nm, the laser setup was equipped with an NIR detector to cover the spectral region up to 1700 nm. The actual absorption spectrum from 800 to 1700 nm was obtained by mapping the area, which means measuring kinetic absorption every 10 nm and subsequently slicing the collected data in equal sections, using an available feature of the software (P900). Directly observable from the transient data is the very low optical density and a rather low signal to noise ratio, yet all relevant information could still be extracted. Variation of the sample concentration did not result in any improvement of the signal. Considering the transient data, the most likely photoproduct is the singly oxidized (Re+)(Re0)–OTA+ with absorption maxima at 405, 780, and 1360 nm. From the spectrum between 400 and 800 nm alone, it is difficult to make a clear statement about the nature of the photoproduct due to the above discussed points. However, the lack of the prominent band at 1140 nm strongly suggests that little or no (Re0)2–OTA2+ was generated in the experiment at hand. Though thermodynamically favored (∆G 0 0 2+ would ET = −0.71 eV), formation of the second charge-separated state (Re )2–OTA require the successful absorption of a second photon which seemed not to happen in the present case. Furthermore, additional deactivation pathways are open for (Re0)2– OTA2+ that are not available to (Re+)(Re0)–OTA+. This will decrease the lifetime of (Re0)2–OTA2+ significantly. Thus, the observation that (Re+)(Re0)–OTA+ is the predominant species generated in the above experiment is not a big surprise. 1 420 nm 790 nm

0.1 0

20

40

60

80

100

time / ns

Figure III.11 Transient absorption decays of the photoproduct of Re2–OTA at 420 nm and 790 nm.

The lifetime of 36 ns extracted from the decays from Figure III.11 resemble those obtained in the reference complex Re2–TAA. Unfortunately, due to the low signal strength and the relatively poor sensitivity of the NIR detector it was not possible to measure a decent transient absorption decay in the infrared region. On account of the short lifetime of (Re+)(Re0)–OTA+, our attention turned towards the use of the flash-quench technique. This approach should significantly increase the charge-separated state lifetime, since the electron will be transported further away

45

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes from the OTA radical cation and recombination will then be a diffusion controlled step.

III.1.10 Oxidative Flash-Quench Experiments The basic principle of the flash-quench experiment is depicted in Scheme III.5 with an exemplary complex Re–OTA. *Re(I)-OTA

hn

intermolecular ET

Q-

Q

Re(I)-OTA

Re(II)-OTA

bimolecular recombination

intramolecular ET

Re(I)-OTA+

Scheme III.5 Principle of flash-quench experiments.

Once in its excited form, the rhenium photosensitizer (*Re(I)), as a strong reductant, is able to reduce an appropriate quencher (Q). Thus, intermolecular electron transfer forms the reduced quencher Q– and oxidized Re(II). This is followed by rapid intramolecular electron transfer from the OTA donor to Re(II). The recombination process resulting in the molecules in their ground-states, Re(I)–OTA and Q, is now diffusion controlled. This will prolong the lifetime of the charge-separated species significantly, typically up to several microseconds. By this approach, we hoped to increase the probability of exciting (Re+)(Re0)–OTA+ a second time with our nanosecond setup. A suitable quencher needed to fulfill two major criteria: First, its reduction by the excited rhenium photosensitizer had to be thermodynamically possible. Second, its reduced form should not display any absorption beyond 350 nm to prevent superposition with bands of oxidized OTA and absorption of laser light. Based on these requirements, 1,2,4,5-tetracyanobenzene (TCNB) was chosen as the quencher for the following flash-quench experiment. Its reduction potential of Ered = −1.04 V vs. Fc+/0 allows for *Re(I) (*Eox = −1.94 V vs. Fc+/0) to reduce it, and TCNB – shows only weak absorption in the spectral range we are interested in. Furthermore, it proved to be very stable in de-oxygenated solution.116 The left part of Figure III.12 shows the transient absorption spectrum of a deoxygenated 10−5 M MeCN solution of Re2–OTA with an excess of TCNB (0.2 M). The spectrum was recorded directly after the excitation pulse in a 200 ns time window.

46

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes a)

normalized optical density

15

mDOD

10 5 0 -5 -10

300

Figure III.12

400 500 600 700 wavelength / nm

800

1

b)

375 nm 430 nm 460 nm 790 nm

0.1

0.01 0

100

200

300

400

time / µs

a) Transient absorption spectrum after 355 nm excitation of a deoxygenated 10−5 M MeCN solution of Re2–OTA with an excess of TCNB; b) Corresponding decays at different wavelengths.

Directly observable is the similarity of the flash-quenched transient absorption spectrum and the one obtained after excitation of the pure complex solution, with bands at 430 nm and 790 nm being again attributed to the singly oxidized Re2–OTA. The two new bands at 375 nm and 460 nm belong to the reduced TCNB, as was demonstrated by Protti et al. who generated TCNB – quantitatively by addition of an excess of NEt3 to a de-oxygenated MeCN solution of TCNB.116 Investigations in the NIR region were impeded by the same reasons already mentioned above, especially the poor sensitivity of the detector. The signals from Figure III.12a) decay on a 400 µs timescale (Figure III.12b)). Alas, despite the prolonged lifetime of Re2–OTA+, even the flash-quench method could not produce any sign of the twofold charge-separated species. Thus, the multitude of deactivation processes open to Re2–OTA2+ could not be influenced to ensure its higher population. In addition, prolongation of the lifetime of Re2–OTA+ also did not result in the desired photoproduct. Consequently, our focus turned towards studying the influence of variations in the excitation power output.

III.1.11 Influence of the Excitation Density on the Photochemistry Generating OTA2+ requires the absorption of two photons upon excitation which is intrinsically more difficult to achieve and also less trivial than simple separation of one electron from one hole. Studies on photon-upconversion have shown that the dependence of two-photon absorption on the excitation density is often of quadratic nature.110,111 By variation of the excitation power, a change in the transient absorption spectrum was anticipated which might indicate the formation of OTA2+ on top of a stronger signal caused by OTA+. It is known from the chemical oxidation of Re2–OTA, as well as from earlier studies on oligotriarylamines that OTA2+ shows increased absorption at

47

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 600 nm compared to OTA+.52,77,96 Figure III.13 shows the transient absorption spectra of Re2–OTA in MeCN solution recorded at high (3.8 mJ, cyan trace) and low (1.1 mJ, blue trace) excitation power as well as the difference spectrum (black trace). To obtain the difference spectrum, the low-energy spectrum was multiplied by a factor of 3.45, the quotient of 3.8/1.1 mJ and then subtracted from the cyan spectrum recorded at an excitation power of 3.8 mJ. A laser density of 1.1 mJ proved to be the lower limit. Decreasing the excitation intensity further impaired the already weak transient signal and the signal-to-noise ratio.

mDDOD

mDOD

6 4 2 0 -2 4 2 0 -2 -4 400

500

600

700

800

wavelength / nm

Figure III.13 Intensity-dependent transient absorption spectra of a de-oxygenated 10−5 M MeCN solution of Re2–OTA. Cyan: 3.8 mJ, blue: 1.1 mJ, and the difference spectrum in black.

Despite the increase in the excitation density to 3.8 mJ, no increased absorption around 600 nm could be detected as was expected for the formation of OTA2+. Though, measurements at different laser powers did not reveal any spectral changes indicative of OTA2+, power-dependent kinetic absorption spectra at 420 nm were recorded to corroborate these results. On the basis of the lack of spectral evidence for OTA2+, a linear dependence of the signal intensity on the laser excitation power P is expected. This expectation is met as can be directly observed from the graph in Figure III.14a).

Figure III.14 Dependence of the signal intensity on the laser power. Left: linear depen+ dency due to formation of (Re+)(Re0)–OTA√ in a single-photon process. Right: non-linear behavior of intensity vs. P because no two-photon process takes place.

48

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes By the provision of TCNB as an electron acceptor we hoped to promote the second charge separation as discussed in Subsection III.1.10. Recombination now being a diffusion controlled step, should increase the possibility of forming Re2–OTA2+ with a detectable concentration and lifetime. However, despite the increased lifetime of Re2– OTA+ in presence of the quencher, still no spectral evidence for OTA2+ was observed (Figure III.15) at 3.1 mJ. Even higher laser powers result in the decomposition of the sample by photodissociation of CO ligands of the photosensitizer.117 mDOD

20 10 0

mDDOD

-10 10 5 0 -5 -10 300

400

500

600

700

800

wavelength / nm

Figure III.15 Intensity-dependent transient absorption spectra obtained from the same solutions as in III.1.10. Cyan trace: 3.1 mJ, blue trace: 1.0 mJ, black: difference spectrum obtained by subtracting the blue trace from the cyan trace.

Even laser intensity dependent measurements could not reveal twofold oxidized oligotriarylamine in the present Re2–OTA triad. As discussed earlier, if generated at all, (Re0)2–OTA2+ will most likely have a very short lifetime due to the manifold of possible deactivation pathways of this high-energy photoproduct. It might in fact be possible that the lifetime of the twofold charge-separated state is less than 10 ns, which is the pulse duration of the laser, making it impossible to detect (Re0)2–OTA2+ with our laser setup. In an attempt to increase the lifetimes of the charge-separated states by elongating the D-A distance, a triad comprised of an anthraquinone (AQ) acceptor, a rhenium photosensitizer, and the OTA donor was synthesized. Another process that may be responsible for the lack of the second charge-separated state is undesired energy transfer between the rhenium photosensitizer in its excited 3 MLCT state and the present oxidized amines in the first charge-separated state. Due to the low-energy absorptions of TAA+ and OTA+ this undesired deactivation process might indeed occur. In order to assess its possibility, luminescence lifetime measurements were performed on de-oxygenated MeCN solutions of Re2–TAA+ and Re2– OTA+ (Figure III.16). If energy transfer from the excited rhenium sensitizer to the oxidized amines occurs, the emission lifetimes are expected to be shorter than for the triads before amine oxidation. Measurements of the non-oxidized species gave luminescence lifetimes of

49

normalized optical density

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 1

0.1

0.01

0.001 0

200

400

600

800

1000

1200

time / ns

Figure III.16 Emission decays of Re2–TAA+ (blue) and Re2–OTA+ (cyan) after excitation of a de-oxygenated MeCN solution at 355 nm.

21 ns for Re2–TAA and 27 ns for Re2–OTA (III.1.8). In the reference experiment on the oxidized species the emission lifetimes are 25 ns for Re2–TAA+ and 27 ns for Re2– OTA+. Within the accuracy of the measurement and the fit of the decays the 3MLCT lifetimes of the non-oxidized species, compared to the data obtained from Re2–TAA and Re2–OTA before amine oxidation are the same. It is therefore concluded that energy transfer between the excited photosensitizer and the oxidized amines is rather inefficient. However, another possibility for unwanted deactivation is reverse electron transfer from the oxidized amines to the excited photosensitizer after absorption of a second photon. As discussed in II.4 this undesired electron transfer reaction can also contribute to not achieving the twofold charge-separate state.

50

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.2 Electron Transfer in an AQ–Re–OTA Triad Based on the insights gained from the studies on Re2–OTA and known photochemistry and kinetics of donor-photosensitizer-acceptor molecules, a structurally simple triad comprised of an anthraquinone acceptor (AQ), a rhenium photosensitizer, and the OTA donor was synthesized. The design of the triad was aimed at accumulating two holes on the OTA and the transferred electrons on the AQ acceptor. It is already known from Nature that quinones are two-electron acceptors. Charge accumulation in this triad does require the twofold excitation of the one rhenium photosensitizer. In the following section efforts on our side for achieving this twofold charge separation will be discussed.

III.2.1 Synthesis of AQ–Re–OTA For the synthesis of AQ–Re–OTA some general building blocks, such as 2,2’bipyridine derivatives and a xylene bridging unit, were needed. The different bipyridine derivatives used for ligand synthesis are 5-bromo-2,2’-bipyridine (Br–bpy) and 5-(tributylstannyl)-2,2’-bipyridine (py–SnBu3). The reaction pathway for their synthesis is depicted in Scheme III.6.

N

ZnBr

+

I

N

Br

a) 78 %

N

N

Br-bpy

Br

b) 58 %

N

N

SnBu3

bpy-SnBu3

Scheme III.6 Synthesis of the bpy building blocks: a) Pd(PPh3)4, THF, RT, 24 h; b) nBuLi, tributyltin chloride, Et2O, RT, over night.

All of these molecules are known in literature and were synthesized accordingly or with slight modifications in the purification steps. Therefore, their synthesis will not be discussed in detail here.118–120 In the AQ–Re–OTA triad, the distance between donor/acceptor and the pyridine/bipyridine coordination sites was elongated by inserting a xylene bridging unit. Being a frequently used spacer in earlier works, its synthesis (Scheme III.7) mostly followed literature procedures. For the first step, the instructions of Freys et al.119 were accurately followed, and for the introduction of the boronic acid, a procedure of So et al. was adapted.121 In this way, the xylene bridging unit could be obtained in an overall yield of 70%. Synthesis of the AQ–Re–OTA triad begun with preparation of an AQ–Re precursor (Scheme III.8) which was later reacted with a py–xy–OTA ligand to yield the final

51

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

a) 93 %

Br

Br

Br

TMS

b) 75 %

B(OH)2

TMS

TMS-xy-B(OH)2

TMS-xy-Br

Scheme III.7 Synthesis of the xylene spacer: a) nBuLi, trimethylsilyl chloride, THF, RT, over night; b) nBuLi, trimethyl borate, THF, RT, over night, 6 M HCl.

AQ–Re–OTA complex. For obtaining the required AQ–xy–bpy ligand, a previously published procedure of our own group was followed.88 O

Br

X

O + (HO)2B

TMS

O

a) 100 %

O

TMS-xy-B(OH)2

AQ-xy-X X = TMS

b) 100 %

X = I + Bu3Sn

N

N

c) 73 % O O O

N

N

OC Re X

AQ-Re-X e) 83 %

OC

CO

d) 91 %

O

N

N

AQ-xy-bpy

X = Cl X = OTf

Scheme III.8 Synthesis of the AQ–Re–OTf precursor: a) Na2CO3, Pd(PPh3)4, toluene/EtOH/H2O, reflux, 1.5 d; b) ICl, DCM/MeCN, 0 °C/RT, over night; c) Pd(PPh3)4, m-xylene, 150 °C, 70 h; d) Re(CO)5Cl, toluene, reflux, over night; e) trifluoromethanesulfonic acid, DCM, RT, 2 h, Et2O.

Suzuki coupling of 2-bromo-9,10-anthraquinone to TMS–xy–B(OH)2, as well as subsequent deprotection of the TMS group with ICl was achieved in essentially quantitative yields for each reaction. Preparation of the final ligand via Stille coupling of AQ–xy– I with bpy–SnBu3 in hot m-xylene required a rather long reaction time of about 70 h, and purification by column chromatography proved challenging. Nevertheless, the desired product could be obtained in sufficient purity and an acceptable yield after recrystallization from pentane. The overall yield was 73%. Complexation of AQ– xy–bpy with Re(CO)5Cl in refluxing toluene and subsequent exchange of the chloro ligand with triflate gave the AQ–Re–OTf precursor. This was then used in a complexation reaction with py–xy–OTA to yield AQ–Re–OTA. The strategy for preparation of py–xy–OTA as well as the final synthetic step to yield AQ–Re–OTA are depicted in Scheme III.9.

52

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

N

B(OH)2

+ (HO)2B

a)

TMS

N

X

99 %

TMS-xy-B(OH)2

py-xy-X X = TMS

b) 97 %

X=I

+

OTA c) 94 %

H3CO O N O

N OC

AQ-Re-OTf

OCH3

N OTf

Re

OC

+

N

OTA

N

CO N d) 47 %

OCH3

H3CO + OTf-

H3CO O

O

N OC OC

N

OCH3

N

OCH3

N Re

N

N

CO

AQ-Re-OTA H3CO

Scheme III.9 Reaction pathway to AQ–Re–OTA: a) Na2CO3,Pd(PPh3)4, toluene/EtOH/H2O, reflux, 1.5 d; b) ICl, DCM/MeCN, 0 °C/RT, over night; c) KOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, 100 °C, 2.5 d; d) MeOH/CHCl3, reflux, 43 h.

For the synthesis of the py–xy–OTA ligand the first two steps were performed in analogy to AQ–xy–I. Suzuki coupling of 4-bromopyridine hydrochloride with TMS–xy– B(OH)2 followed by TMS to iodine exchange by means of ICl gave py–xy–I. In a subsequent Pd-catalyzed Buchwald-Hartwig coupling with OTA the final ligand was obtained in an overall yield of 90%. AQ–Re–OTA was then synthesized by complexation of the AQ–Re–OTf precursor and py–xy–OTA. After column chromatographic purification and recrystallization from Et2O the desired complex was received as the triflate salt. Prior to performing electrochemical and spectroscopic measurements the complex was fully characterized by 1H NMR, HRMS, and elemental analysis.

53

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.2.2 Spectroscopical Characterization of the Triad The FT-IR spectra of AQ–Re–OTA (blue trace) and Re–ref (black trace) are both displayed in Figure III.17 for direct comparison. As already observed for the rhenium triads in Subsection III.1 the most prominent features around 2000 cm–1 are attributed to totally symmetric and asymmetric stretching vibrations of the three carbonyl ligands.103 The additional band at 1670 cm–1, as a typical frequency for quinone vibrations, is most likely due to a stretching of the anthraquinone-related carbonyls. Furthermore, it is neither visible in the infrared spectrum of Re–ref nor in the ones of Re2–OTA or Re2–TAA. 100

transmission / %

90 80 70 60 50 40 1000

1500

2000

wavenumbers / cm

500

-1

Figure III.17 Near infrared spectra of Re–ref (black) and the new triad AQ–Re–OTA (blue).

The fingerprint region of AQ–Re–OTA between 400 and 1400 cm–1 shows great resemblance to Re2–OTA and Re2–TAA since it contains the same functional groups such as N-heterocycles, methoxy groups, and arylamines. An increase in absorbance is also observed in the optical absorption spectrum of a 10−5 M MeCN solution of AQ–Re– OTA. Compared to Re–ref and Re2–OTA, it exhibits distinctly stronger absorption, possibly due to its more extended π-system. At 355 nm the extinction coefficient of 49500 M–1cm–1 is relatively large. 80

0.8 0.6

40

0.4 20 0.2

relative intensity a.u.

extinction coefficient / 103 M-1cm-1

1.0 60

0

0 300

400

500

600

700

800

wavelength / nm

Figure III.18 Optical absorption (left) and steady-state emission spectrum (right) of 10−5 M MeCN solutions of AQ–Re–OTA (blue) and Re–ref (black). The emission spectra were recorded after excitation of the de-oxygenated complexes at 355 nm.

54

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes The left part of Figure III.18 shows that the absorption of AQ–Re–OTA increases rapidly below 400 nm. The 1MLCT transition around 360 nm is not discernible anymore, but a shoulder at 425 nm probably attributable to intramolecular charge-transfer between the attached electron-donating substituents and the carbonyls on AQ is visible.122 The band structure between 250 and 330 nm seen here is a superposition of π − π∗ -transitions of the rhenium photosensitizer, OTA, and the newly introduced AQ acceptor.80,86,122 As in Re2–OTA, the MLCT emission is completely quenched in AQ–Re–OTA. Based on the results obtained from Re2–OTA it is therefore likely that photoinduced electron transfer is taking place in AQ–Re–OTA. With a triplet energy of 2.72 eV for AQ,123 energy transfer from the excited rhenium (ET = 2.82 eV) to anthraquinone would be exergonic by 0.1 eV, however, we reasoned that the very exergonic and therefore presumably very fast oxidation of OTA will be the dominating process and allow for the desired photoproducts to be built. Conclusive evidence for photoinduced ET will be presented below.

III.2.3 Electrochemical Study of AQ–Re–OTA and Establishment of the Energy-Level Diagram Concerning the electrochemical properties of AQ–Re–OTA a similar behavior as for Re2–OTA was expected. Figure III.19 shows the recorded cyclic voltammograms of a 1 mM solution of the triad in MeCN. 0.1 M TBAPF6 was again used as the supporting electrolyte and Me10Fc as internal standard, and Re–ref (black trace) is also shown to facilitate comparison of redox potentials. In this case, as for Re2–TAA and Re–ref, separate recording of oxidation and reduction provided superior results.

Figure III.19 Cyclic voltammogram of AQ–Re–OTA (blue) as a 1 mM solution in MeCN. The waves at −0.5 V correspond to the internal reference Me10Fc. Re–ref (black) is depicted for direct comparison.

The values of 0.95 V for the rhenium oxidation and −1.8 V for the bpy reduction closely resemble those obtained in the Re2–OTA triad. Here, the metal center is oxi-

55

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes dized easier than in the Re–ref complex for the same reason as discussed in Subsection III.1.5, namely the covalent attachment of the very electron rich OTA donor. Oxidation of the OTA donor occurs at equally low potentials, −0.05 V for the first and 0.21 V for the second oxidation, as was found in Re2–OTA and other systems containing oligotriarylamines.77,96 With the attachment of the anthraquinone acceptor, two new reduction waves have to be considered. At −1.29 V vs. Fc+/0 AQ is reduced to AQ · and the second reduction to AQ2– around −1.8 V overlaps with the bpy reduction. The additional wave at −1.58 V is assumed to be due to an electrochemical decomposition product.

Scheme III.10 Energy-level diagram for the theoretical formation of AQ2––Re–OTA2+.

The energy-level diagram depicted in Scheme III.10 clearly shows that the thermodynamical driving force for the generation of the first charge-separated state AQ––Re– OTA+ is −1.58 eV. Considering only the ground state redox potentials extracted from the cyclic voltammogram, formation of AQ2––Re–OTA2+ after two-photon absorption is exergonic by 0.81 eV. However, the excited rhenium photosensitizer (*Eox = −1.87 V vs. Fc+/0) as well as its reduced species can not reduce AQ (E2.Red = −1.8 V) a second time. Former studies on the electrochemical behavior of quinones have shown that upon addition of acids or strongly hydrogen bond-donating solvents the two one-electron reduction waves for the first and second reduction of the quinone merge forming a new two-electron reduction wave at more positive potentials.90,93,124 This circumstance is in fact crucial for the formation of the second charge-separated state in AQ–Re–OTA due to the above mentioned reasons. Addition of acid will shift the potential of the second AQ reduction to less negative values probably around −1.2 V, allowing for the rhenium photosensitizer to reduce it a second time. The expected photoproduct

56

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes hereby is then AQH2–Re–OTA2+.

III.2.4 Influence of Acids on the OTA Donor Prior to transient experiments on the triad, OTA–ref was titrated with chloroacetic acid and triflic acid (pKa = 0.7 in MeCN).125,126 Although the addition of acid is necessary to protonate AQ and enable twofold charge separation, protonation of the OTA donor is not desirable. Triarylamines in general exhibit low basicity due to their propeller-like shape. From the crystal structure in Figure III.1 it is known that OTA exhibits a similar propeller-like structure. Therefore, protonation by chloroacetic acid is not expected.

Figure III.20 UV/Vis absorption spectra of 10−5 M MeCN solutions of pure OTA–ref (black), after addition of an excess of chloroacetic acid (blue), and triflic acid (cyan).

However, titration of OTA–ref with the very strong acid triflic acid was performed to clarify if the oligotriarylamine could be protonated at all. Figure III.20 shows the UVVis spectroscopical signatures of pure OTA–ref (black), with an excess of chloroacetic acid (blue), and triflic acid (cyan), respectively. Observable from the graphs is that chloroacetic acid, later used in the transient experiments, has negligible effects on the absorption spectrum of OTA–ref. For triflic acid, on the other hand, absorption bands at 420 and 1280 nm appear. These bands, though still very weak compared to the initial OTA–ref signature, closely resemble the spectral signatures obtained by chemical oxidation of the OTA donor in AQ–Re–OTA with Cu(ClO4)2 as depicted in Figure III.21. Chemical formation of AQ–Re–OTA+ by means of Cu(ClO4)2 does not influence spectral features of the AQ acceptor so that the absorption spectrum is more or less identical to the one obtained by oxidation of Re2–OTA. Local maxima at 426, 733, and 1300 nm are attributed to OTA+, whereas new absorptions arise at 590 and 1127 nm with a concomitant disappearance of the band at 426 nm when the amine is further oxidized to OTA2+.

57

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

Figure III.21 Stepwise chemical oxidation of a 10−5 M MeCN solution of AQ–Re–OTA with Cu(ClO4)2. Blue trace = OTA+, cyan trace = OTA2+.

Considering the non-redox active triflic acid used here, it is surprising to find features of oxidized OTA upon acid addition.127 Kantner et al., however, found the apparent oxidation of triphenylmethane by triflic acid was rather due to elimination of one benzene followed by hydride transfer from an unreacted triphenylmethane. Therefore, making it a disproportionation not an oxidation.128 Despite the lack of a transferable hydrogen atom in OTA–ref, the donor molecule is very electron-rich. It might undergo similar mechanisms as described above resulting in the seemingly oxidized species observable in the spectrum. However, chloroacetic acid, not triflic acid, was the strongest acid used as a proton source in the following transient experiments. This less strong acid induced no spectral changes in the preliminary test. Hence, protonation of the OTA donor by chloroacetic acid seems unlikely in MeCN.

III.2.5 Influence of Hydrogen-Bond Strength on the Photoproducts As discussed above, reduced AQ needs to be protonated for the second charge separation to occur. Experiments were therefore conducted with de-oxygenated MeCN solutions of the pure complex, with addition of benzoic acid (pKa = 21.5 in MeCN),129 and addition of chloroacetic acid (pKa = 18.8 in MeCN)130 to explore the influence of acids of different strengths on the nature of the generated photoproducts. In principle, already benzoic acid should be able to protonate AQ– rendering stronger acids unnecessary.93 Kinetic emission measurements on the de-oxygenated MeCN solutions of AQ–Re– OTA with and without acids (Figure III.22) showed complete quenching of the initial rhenium emission, as was already seen in the steady state emission experiment. The lifetimes extracted from single exponential fits were shorter than 10 ns which is the instrumental limit. For the spectra showed in Figure III.23 excitation occurred with a 10 ns laser pulse at 355 nm and detection in a 200 ns time window (100 ns with benzoic

58

normalized optical density

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 1

0.1 0

5

10

15

20

time / ns

Figure III.22 Luminescence decays of 10−5 M MeCN solutions of pure AQ–Re–OTA (black), with 0.2 M benzoic acid (blue), and 0.2 M chloroacetic acid (cyan).

acid) directly after the pulse. For the triad in pure MeCN absorption bands at 427, 564, and 1290 nm, as well as a bleach at 330 nm can be observed. The bleach and the 564 nm absorption can be attributed to AQ –, whereas the band at 427 nm and especially the very broad band in the NIR peaking at 1290 nm are assigned to singly oxidized OTA.90,93,96,113 These results are indeed expected for AQ––Re–OTA+ in pure MeCN since the absence of a proton source or a Lewis acid should not allow for the twofold charge separation to occur.

Figure III.23 Transient absorption spectra of de-oxygenated MeCN solutions of pure AQ–Re–OTA (black), with 0.2 M benzoic acid (blue), and 0.2 M chloroacetic acid (cyan) after 355 nm excitation.

In the case of added benzoic acid the bleach at 330 nm still persists. Wightman et al. showed that benzoic acid was able to protonate AQ – in DMF solution yielding the hydroquinone form of AQ (AQH2) which exhibited spectral features at 384 and 420 nm. The same was observed by Babaei et al.90,93 Twofold oxidation of OTA would lead to the disappearance of the band at 427 nm with a rise of absorption at 590 nm which was not observed in the blue spectrum. The band at 423 nm could have been considered a sign of hydroquinone formation and corresponding twofold oxidized OTA, was it not for the lack of the very prominent band at 1127 nm attributed to OTA2+. Furthermore, the transient spectrum shows a band at 512 nm which can be assigned to AQH, the semiquinone form of AQ. In summary, it can be stated that the most likely photoproduct in presence of benzoic acid is the singly charge-separated

59

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes state stabilized by hydrogen-bonding via the acid. The transient absorption spectrum of AQ–Re–OTA with an excess of chloroacetic acid (cyan trace in Figure III.23) shows slight differences to the two previously discussed ones except for the still present bleach at 330 nm. A band at 409 nm is clearly visible which is blue-shifted by 14 nm compared to the data with benzoic acid, and there is no additional band around 500 nm anymore, yet it might be shifted further into the blue and be hidden under the band at 409 nm. Though twofold reduced and protonated AQ (AQH2) does not exhibit spectral features beyond 420 nm, formation of twofold oxidized OTA can not be assumed due to the lack of the band at 1127 nm. It is therefore concluded that the main photoproduct in presence of chloroacetic acid is the first charge-separated state consisting of OTA+ and singly reduced and protonated AQ.

III.2.6 Kinetics of the Photogenerated Products The transient spectra show three different photoproducts depending on the acid strength. The simple charge-separated state AQ––Re–OTA+ in pure MeCN, the hydrogen-bonded AQ––Re–OTA+ in presence of benzoic acid, and AQH–Re–OTA+ with an excess of chloroacetic acid. Therefore, differences in the lifetimes are expected. Earlier studies on hydrogen-bonding effects of different acids and hydrogenbond donating solvents have shown that upon their addition, the lifetimes of the charge separated-states with quinone acceptors can increase significantly.88,131 In Figure III.24 the decays of the transient absorption signals of AQ–Re–OTA in pure MeCN are depicted in the left graph. Fitting of the decays gives an average lifetime of 205 ns for the generated photoproduct AQ––Re–OTA+. Upon addition of chloroacetic acid the lifetime of the charge-separated state (AQH–Re–OTA+) is indeed prolonged as can be seen on the rightmost side of Figure III.24 and single exponential fitting gives τ = 3.7 µs. The data obtained from AQ–Re–OTA in de-oxygenated MeCN with 0.2 M benzoic acid (Figure III.24 b), on the other hand, clearly shows biexponential decays with τ = 101 ns and τ = 2.3 µs. This presence of two different species, as indicated by the biexponentiel decays, corroborates the interpretation of the transient absorption spectrum. In presence of benzoic acid, AQ––Re–OTA+ and its hydrogen-bonded equivalent seem to be in equilibrium. Since the lifetimes resemble that of AQ–Re–OTA in pure MeCN (205 vs 101 ns) and that observed in presence of 0.2 M chloroacetic acid (3.7 vs 2.3 µs), simultaneous appearance of photoproducts with AQ– and hydrogen-bonded AQ– is suggested. Considering the pka values for benzoic acid and chloroacetic acid, the latter is the stronger acid and therefore able to fully protonate AQ, generating only the

60

normalized optical density

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 1

a)

1

390 nm 420 nm 565 nm 790 nm

0.1

b)

1

420 nm 670 nm 790 nm

0.1

0.1

0.01 2

6

4

8

10

12

14

410 nm 675 nm 790 nm 1240 nm

0.01

0.01 0

c)

0

2

time / µs

6

4

8

10

12

14

0

time / µs

2

4

6

8

10

12

14

time / µs

Figure III.24 Transient absorption decays of photoproducts generated after excitation with a 10 ns laser pulse at 355 nm of a de-oxygenated MeCN solution of AQ–Re–OTA (a), with 0.2 M benzoic acid (b), and 0.2 M chloroacetic acid (c).

more stable semiquinone form.

III.2.7 Dependence of the Transient Spectral Absorption on the Excitation Power

mDDOD

mDOD

Formation of the twofold charge-separated state AQH2–Re–OTA2+ is again a twophoton process and therefore expected to be quadratically dependent on the laser excitation density. To test this hypothesis on AQ–Re–OTA, transient spectra of a de-oxygenated MeCN solution with 0.2 M chloroacetic acid were recorded at high (3.4 mJ) and low (1.1 mJ) excitation powers. In Figure III.25 the obtained spectra after excitation at 355 nm with 3.4 mJ (cyan) and 1.1 mJ (blue) per pulse are depicted. The difference spectrum (black) was obtained by multiplication of the cyan spectrum with a factor of 3.09 (3.4 mJ/1.1 mJ) and subtraction from the blue spectrum. 20 10 0 -10 -20 -30 4 2 0 -2 -4 300

400

500

600

700

800

wavelength / nm

Figure III.25 Transient absorption spectra of a 10−5 M MeCN solution of AQ–Re–OTA with 0.2 M chloroacetic acid at 3.4 mJ (cyan trace) and 1.1 mJ (blue trace) laser power. The blue spectrum was multiplied by a factor of 3.09 and subtracted from the cyan spectrum to give the difference spectrum in black.

Generation of AQH2–Re–OTA2+ should lead to a rise of absorption around 600 nm

61

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes as was earlier established by chemical oxidation of the complex with Cu(ClO4)2. In addition, it should show positive absorption at 400 nm, indicative of formation of the hydroquinone form of AQ (AQH2). Alas, no change was observed in the transient spectrum after increase of the excitation power.

III.3 Summary and Conclusions on Photoinduced Electron Transfer in Rhenium-OTA-Complexes The preceding two sections gave an insight in our efforts on creating a wholly molecular, nanoparticle-free donor-acceptor system with rhenium polypyridyl complexes as the photosensitizers. Two new triads, Re2–OTA and AQ–Re–OTA were synthesized, characterized, and studied in terms of their ability to undergo twofold charge separation. The main advantage of Re2–OTA is the presence of two rhenium photosensitizers. This should allow for the simultaneous absorption of two photons within one single laser pulse, provided the excitation laser power is high enough. In AQ–Re– OTA, though having only one photosensitizer, it is possible to separate the charges over a longer distance. This should increase the lifetime of the first charge-separated state and the possibility of a second excitation before charge recombination occurs. It is known from earlier studies that the formation of the first charge-separated state in a similar AQ-Ru-TAA triad occurs within ≈ 200 ps.132 Since OTA has proven to be an even better electron donor, formation of AQ––Re–OTA+ is expected to lie in the same time range or it could be even faster. It was therefore concluded that twofold charge separation should be possible with relatively long (10 ns) and sufficiently strong (>3 mJ) pulses. Kinetic absorption experiments on Re2–OTA and AQ–Re–OTA indeed showed instant formation of the photoproducts within the duration of the laser pulse (10 ns). In contrast to AQ–Re–OTA with a lifetime of 205 ns for the singly charge-separated state, Re2–OTA+ decays to its ground state in 36 ns. A conjectural value for the close proximity of the OTA donor and rhenium acceptors. Furthermore, AQ–Re–OTA undergoes PCET in presence of acids which increased the lifetime of the generated AQH– Re–OTA+ to 3.7 µs. In addition, transient absorption spectra of pure Re2–OTA and AQ–Re–OTA with chloroacetic acid were recorded at different laser excitation powers since the formation of the twofold charge-separated state, being a two-photon process, should exhibit quadratic dependence on the excitation intensity. Alas, even at the highest possible laser power there were no spectral features observable that could be unambiguously attributed to formation of AQH2 or OTA2+. Higher powers than ≈4 mJ led to decomposition of the sample solutions, especially in presence of acids.

62

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes A major challenge of creating a molecular system capable of twofold charge separation is to minimize the multitude of possible unwanted deactivation pathways that are open to the second charge-separated state enabling its fast decay. In the investigated complexes the lifetime of the second charge-separated state, compared to the first, might be very short and its population very small, if it is generated at all. These circumstances make its detection difficult. The major problem from a spectroscopic point of view was probably the choice of rhenium polypyridyl complexes as photosensitizers. Though synthetically more easily accessible than their ruthenium analogs, because ligand synthesis could be split in a pyridine and a bipyridine-part, they are prone to carbonyl ligand-loss. Furthermore, the short excitation wavelength (355 nm) required to excite the photosensitizers will also result, at least partly, in the excitation of AQ and OTA which both show significant absorption at this wavelength. In addition, once OTA+ is generated as a photoproduct, a second pulse will excite the oxidized OTA since it also exhibits absorption around 355 nm. These circumstances might interfere with the formation of the second charge-separated state. Lastly, the most sobering insight we gained from the experiments on the rhenium triads was that with around 4 mJ the excitation density was simply not high enough to initiate twofold charge separation. A 355 nm laser pulse of 4 mJ provides 7·1015 photons and only one third of the molecules of a 10 µM solution in the cuvette can be excited by the pulse, i.e. 6·1015 molecules. As a result, a 4 mJ laser pulse provides roughly one photon per molecule which hampers the possibility of generating a twofold charge-separated state in our rhenium complexes.

63

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.4 Photoinduced Electron Transfer in an AQ–Ru–OTA Triad The following section will focus on our efforts to achieve accumulative electron transfer in an AQ–Ru–OTA triad. Being the ruthenium analog of the rhenium triad discussed in III.2 we hope to eradicate some flaws encountered in the rhenium complex. First, the ruthenium MLCT band should not be superimposed by absorptions from OTA and AQ, allowing for the exclusive excitation of the ruthenium sensitizer. Second, the laser setup for excitation at 532 nm should allow us to measure with higher intensities gaining a higher photon flux, hopefully resulting in two-photon absorption. In addition to the laser experiments, continuous irradiation experiments were performed monitoring the charge separation behavior over a longer time period.

III.4.1 Synthesis of the AQ–Ru–OTA Triad and a Reference Dyad In order to prepare the triad yet another bpy derivative, Br–bpy–Br, was synthesized as depicted in Scheme III.11 by adopting a previously published protocol.120 Br–bpy–Br was hereby obtained via Stille-coupling in 80% yield as a beige solid. The AQ acceptor was attached to Br–bpy–Br via a xylene bridging unit. For that, AQ–xy–I was reacted in a Suzuki-coupling with bis-pinacolato diboron to yield AQ–xy–B-ester.

I

2

a)

Br N

Br

N

80 %

N

Br

Br-bpy-Br

I

O

b)

O

O BO

91 % O

AQ-xy-I

O

AQ-xy-B-ester

Scheme III.11 Synthesis of Br–bpy–Br (top): a) hexa-n-butylditin, Pd(PPh3)4, m-xylene, 180 °C, 3 d; Synthesis of AQ–xy–B-ester (bottom): b) bis-pinacolato diboron, KOAc, PdCl2(PPh3)2, DMF, 100 °C, 24 h.

As depicted in Scheme III.12, AQ–xy–B-ester was then coupled Pd-catalyzed to Br– bpy–Br. The resulting product AQ–xy–bpy–Br precipitated from the hot reaction mixture due to its very low solubility and was directly filtered and washed with MeOH. After column chromatography on silica with an eluent mixture of DCM and 1% NEt3, pure AQ–xy–bpy–Br was obtained as a yellow solid in 92% yield.

64

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

O BO

O

+ Br

N

N

O

a)

Br

N

92 %

Br

AQ-xy-bpy-Br

O O AQ-xy-B-ester

N

Br-bpy-Br O

X = TMS

N

c)

b)

X

N

73 %

AQ-xy-bpy-xy-X

O

66 % X=I

+

H3CO

OTA

d) 50 %

O N O

N

N

OCH3

N

OCH3

N

AQ-bpy-OTA

H3CO 2+ 2PF6-

H3CO

N

OCH3 e)

O N

N

68 %

Ru N

N

O

N

N

N

N

OCH3

AQ-Ru-OTA H3CO

Scheme III.12

Synthesis of the AQ–Ru–OTA triad: a) Na2CO3, Pd(PPh3)4, THF/H2O, reflux, over night; b) TMS–xy–B(OH)2, Na2CO3, Pd(PPh3)4, toluene/EtOH/H2O, reflux, 2 d; c) ICl, DCM/MeCN, 0 °C/RT, over night; d) NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, reflux, 27 h; e) Ru(bpy)2Cl2, CHCl3/EtOH, reflux, 2 d.

To elongate the D-A distance further, another xylene bridge was introduced via Pd-catalyzed Suzuki-coupling of TMS–xy–B(OH)2 to AQ–xy–bpy–Br. Subsequent TMS exchange by I gave AQ–xy–bpy–xy–I in 66% yield as a yellow solid. This was then coupled to OTA to give the ligand AQ–bpy–OTA as a brown solid in 50% yield after column chromatographical purification and recrystallization from hexane. Synthesis of the final complex AQ–Ru–OTA was achieved by refluxing the ligand AQ–bpy–OTA and Ru–pre in a de-oxygenated CHCl3/EtOH mixture for 2 d. The complex was then purified by column chromatography with an eluent mixture of acetone/H2O/sat. KNO3 solution (90:10:1). After concentration of the product fraction, the final complex was precipitated as the PF6 salt by dropwise addition of a saturated KPF6-solution. Filtration, washing with water, and drying yielded the complex as a dark brown solid in 68%. The overall yield for the AQ–Ru–OTA

65

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes complex as described here was 15%. The ruthenium precursor Ru(bpy)2Cl2 (Ru–pre) needed for the preparation of the desired complex AQ–Ru–OTA was synthesized by refluxing RuCl3 · 3 H2O with 2 eq. 2,2’-bipyridine, and lithium chloride under nitrogen atmosphere. The resulting product was precipitated from the cool reaction mixture by dropwise addition of acetone. Filtration and excessive washing with acetone, H2O, and Et2O gave Ru–pre as a dark violet solid.

N

N

Br

+

a) TMS

bpy-Br

B(OH)2

99 %

N

TMS-xy-B(OH)2

bpy-xy-TMS

H3CO

b) 97 % N

N

N

OCH3 c)

N

99 %

bpy-xy-OTA

TMS

N

N

N

I

N

bpy-xy-I

OCH3

H3CO

2+ 2PF6-

H3CO d) 65 %

N N

N

N

Ru-OTA

OCH3

N

OCH3

N

Ru N N

N

H3CO

Figure III.26 Synthetic route for preparation of Ru–OTA: a) Na2CO3, Pd(PPh3)4, H2O/EtOH/toluene, 90 °C, over night; b) ICl, DCM, 0 °C/RT, over night; c) NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, 110 °C, 24 h; e) Ru(bpy)2Cl2, CHCl3/EtOH, reflux, 21 h.

In order to evaluate the influence and relevance of the AQ acceptor in continuous irradiation experiments a reference complex was synthesized. In this Ru–OTA dyad the AQ unit was simply omitted. Figure III.26 depicts the synthetic route followed for its preparation. Starting point hereby was Br–bpy which was reacted in a Pdcatalyzed Suzuki-coupling with TMS–xy–B(OH)2 to give bpy–xy–TMS in essentially quantitative yield. Deprotection of bpy–xy–TMS by addition of ICl gave bpy–xy–I which was then coupled to OTA. This gave the final ligand bpy–xy–OTA as a brown solid in 99% yield. Complexation of bpy–xy–OTA with Ru–pre gave the reference

66

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes dyad Ru–OTA with an overall yield of 62% following the route outlined above. As another reference complex [Ru(bpy)3](PF6)2 was used. This is a standard reference which had been prepared in our group in sufficient amounts and was used in this work as received. It will further be denoted Ru–ref.

III.4.2 Optical Spectroscopic and Electrochemical Characterization Absorption spectra of Ru–ref, Ru–OTA and AQ–Ru–OTA (Figure III.27) show an inπ* transition at 287 nm as well as tense band assigned to a ligand-centered π a broad band around 450 nm for the 1MLCT (dπ(Ru) π*(bpy)) transition.36 The intense shoulder at 340 nm in Ru–OTA (cyan trace in Figure III.27) is attributed to an OTA absorption band since it is not present in Ru–ref. The AQ–Ru–OTA triad shows an additional plateau at 320 nm which is not visible in the dyad and can thus be interpreted to be an absorption of the AQ acceptor.36,86,122 100

0.8 60

0.6

40

0.4

20

0.2

relative intensity a.u.

extinction coefficient / 103 M-1cm-1

1.0 80

0

0 300

400

500

600

700

800

wavelength / nm

Figure III.27 UV-Vis absorption (solid line) and steady state emission spectra (dashed line) of the triad (blue), dyad (cyan), and Ru–ref (black). 10−5 M MeCN solutions were used for the absorption measurements. Steady state emission was measured of de-oxygenated MeCN solutions, the spectra were corrected for differences in the absorption at 450 nm, and normalized to the Ru–ref emission.

As can be observed on the right side of Figure III.27 the excited state emission of the ruthenium photosensitizer in the triad is quenched by 96%. In the dyad the original luminescence is quenched by 98%. This strong quenching of the photosensitizer MLCT luminescence is indicating a non-radiative deactivation process. In the triad it is most likely electron transfer from the OTA donor to the AQ acceptor via the ruthenium sensitizer. In the dyad the electron will be transferred from OTA to the ruthenium photosensitizer which acts as the electron acceptor in this case. Furthermore, the emission maximum is red-shifted from 603 nm in Ru–ref to 612 nm in the triad. In the dyad on the other hand, the emission maximum is again at 603 nm. In order to evaluate the thermodynamic possibility for electron transfer, the redox

67

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes potentials of the synthesized complexes were determined via cyclic voltammetry. Figure III.28 depicts the voltammograms of Ru–ref (top, black), Ru–OTA (middle, blue), and AQ–Ru–OTA (bottom, cyan) measured in Ar-purged MeCN solutions. The scan rate was 0.1 V/s and 0.1 M TBAF6 was used as supporting electrolyte. The waves marked by asterisks correspond to the added internal reference. Fc+/0 was used for Ru–ref, whereas AQ–Ru–OTA was measured in presence of Me10Fc. The sharp peak at -2 V in the cyan trace might be due to unwanted reaction at the electrode, however it does not interfere with the assignment of the redox potentials in the triad.

current / µA

40 20 0 -20 -40

*

10 5 0 -5 -10 20 0 -20 -40

* -2

-1

0

1

potential / V vs. Fc+/0

Figure III.28 Cyclic voltammograms of Ru–ref (black), Ru–OTA (blue) and AQ–Ru– OTA (cyan) in MeCN. The waves marked by asterisks correspond to the internal references Fc+/0 for Ru–ref and Me10Fc for the triad.

The redox potentials were extracted from the data in Figure III.28 and summarized in Table III.3. Directly observable is the conformity of the potentials for the ruthenium oxidation and bpy reductions in Ru–ref, Ru–OTA, and AQ–Ru–OTA. In the case of the studied ruthenium complexes, the photosensitizer-based oxidation potentials are not influenced by the attached electron rich amine. In contrast to AQ–Re–OTA, where the donor and acceptor ligands are two separate units, here the amine donor is part of a rigid ligand, also containing the AQ acceptor. The study of a very similar complex previously synthesized in our group, AQ–Ru–TAA, also showed a lack of shift for the ruthenium redox potentials.113 The oxidation potentials for the stepwise oxidation of OTA to OTA3+ are in good agreement with the literature values found for OTA-containing organic molecules or transition metal complexes.77,96 Considering the first two reversible oxidations for OTA at −0.06 V and 0.20 V, these very low potentials should facilitate the generation of the second charge-separated state of AQ–Ru–OTA. Reduction of AQ occurs at −1.26 V vs. Fc+/0 in good agreement with literature, which gives the first reduction of AQ at −1.20 V vs. Fc+/0 in DMF.90 The second reduction to form AQ2– usually takes place at a potential of −1.9 V. This, however, can not be observed in AQ–Ru–OTA, possibly due to superposition with the bpy reductions. With the help of the cyclic voltammetry data, an energy level diagram (Figure III.29)

68

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes Table III.3 Redox potentials in Ar-purged MeCN solutions. All values are given in V vs. Fc+/0.

redox couple

Ru–ref /V

Ru–OTA /V

AQ–Ru–OTA /V

Ru3+/2+ bpy0/– bpy0/– bpy0/– OTA+/0 OTA2+/+ OTA3+/2+ AQ0/–

0.90 −1.72 −1.92 −2.17

0.94 −1.68 −1.88 −2.11 −0.02 0.24 0.68

0.91 −1.66 −1.92 −2.16 −0.06 0.20 0.64 −1.26

depicting all possible electron transfer steps for the formation of the desired photoproduct AQ2––Ru–OTA2+ can be constructed.

Figure III.29 Energy level diagram for AQ–Ru–OTA.

Formation of the first charge separated state AQ––Ru–OTA+ is exergonic by 0.92 eV. As for AQ2––Ru–OTA2+, ∆GET for the second charge separation after absorption of a second photon is close to zero. The reduced ruthenium photosensitizer (Ered = −1.66 V) will not be able to reduce AQ (Ered = −1.9 V) a second time. Therefore, the addition of Brønsted or Lewis acids is necessary. In their presence, the second reduction potential will be shifted to less negative values, thus, in principle enabling a second charge transfer.90,93

69

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.4.3 Kinetic Emission Behavior

normalized optical density

Studies on the kinetic behavior of the dyad and triad were performed on deoxygenated MeCN solutions. As a reference, Ru–ref was investigated as well. In absence of oxygen, Ru–ref has an emission lifetime of 834 ns in solution. 1

0.1

0.01 0

200 400 600 800 1000 1200 1400 1600 1800 2000 time / ns

Figure III.30 Normalized emission decays of Ru–ref (black), Ru–OTA (blue), and AQ– Ru–OTA (cyan). The decays were recorded on 10−5 M de-oxygenated MeCN solutions.

In agreement with the steady state emission behavior, the lifetime of the ruthenium luminescence is completely quenched (τ < 10 ns). The double exponential decay in Ru–OTA can be explained by highly luminescent trace impurities of [Ru(bpy)3](PF6)2 (Ru–ref); the lifetime of the longer-living species (τ = 899 ns) is very similar to the one of free Ru–ref. Also AQ–Ru–OTA gives a double exponential luminescence decay, indicating the presence of two luminescent species. However, the longer luminescent species with τ = 495 ns is probably not free [Ru(bpy)3](PF6)2, but another highly luminescent impurity not observable in the product characterization. Conclusive evidence for electron transfer in the studied ruthenium complexes will be given below.

III.4.4 Chemical Oxidation of OTA In order to facilitate assignment of the bands corresponding to oxidized species of OTA in the following transient laser experiments, chemical oxidation of OTA was performed. AQ–Ru–OTA, as a 10−5 M solution in MeCN, was oxidized with incremental amounts of Cu(ClO4)2. Successive oxidation gave the blue spectrum in Figure III.31 for OTA+ after the addition of 1 eq Cu2+ and the cyan spectrum for OTA2+ after 2 eq. The spectra were recorded as difference spectra. Thereby, directly subtracting the ground state absorption features of the complex and allowing for direct comparison with the transient absorption spectra. Furthermore, the spectra show two isosbestic points at 478 and 1285 nm. At those wavelengths, OTA+ and OTA2+ show identical

70

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.6

DOD

0.4 0.2 0 -0.2 400

600

800 1000 1200 1400 wavelength / nm

1600

Figure III.31 Difference spectra for the chemical oxidation of OTA to OTA+ (blue) and OTA2+ (cyan) obtained by addition of incremental amounts of Cu(ClO4)2.

absorbances which will later help with the normalization of the transient absorption spectra.

III.4.5 Time-Resolved Absorption Studies on the Ru–OTA Dyad Pure MeCN Solution

Before the investigation of the AQ–Ru–OTA triad, the photoinduced electron transfer behavior of the reference dyad was studied. In Ru–OTA, photoinduced accumulation of two holes on the OTA donor moiety should not occur in pure MeCN solution. OTA would be able to carry two holes upon successive oxidation. The photosensitizer, however, cannot accomodate two electrons. In Figure III.32 the expected transient absorption spectrum of the dyad in a 10−5 M solution in de-oxygenated MeCN is depicted. The sharp spike at 532 nm is due to scattered light from the laser beam. normalized optical density

2

mDOD

1 0 -1 -2 300

400

600 500 wavelength / nm

700

800

1.0

400 nm 650 nm

0.8 0.6 0.4 0.2 0 -50

0 time / ns

50

100

Figure III.32 Transient absorption spectrum (left) and corresponding transient absorption kinetics (right) of the Ru–OTA dyad in a de-oxygenated 10−5 M MeCN solution.

The band at 400 nm indicates the presence of OTA+. Together with the plateau-like signature starting at 600 nm on to lower energies, also observed for (Re+)(Re0)–OTA+, this points towards Ru+–OTA+ as the main photoproduct. Due to the rather low signal-to-noise ratio, measurement in the NIR region was not possible. Kinetic ab-

71

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes sorption measurements at 400 nm and 650 nm showed that Ru+–OTA+ has a lifetime of only 10 ns, which is the detection limit of the nanosecond setup. Due to the close proximity of the OTA electron donor and the ruthenium photosensitizer, acting as the electron acceptor, this very short lifetime of the first charge-separated state is not surprising. Flash Quenching with TCNB and Intensity Dependence

In order to achieve a longer living charge-separated state and possibly hole accumulation in the dyad a suitable electron acceptor needs to be present. The ruthenium polypyridyl photosensitizer cannot accept two electrons and the sole charge accumulating site in this system is OTA. One important prerequisite for successful hole accumulation in this system is a long enough lifetime of Ru+–OTA+ to allow for the absorption of a second photon within the 10 ns laser pulse. In an intermolecular electron transfer experiment between TCNB and the dyad, the back electron transfer would be diffusion controlled. Therefore, provision of an additional electron acceptor, such as TCNB, should lead to the necessary prolongation of the first separated state lifetime. The basic principle of an oxidative flash quenching experiment was already introduced in Section III.1.10 and will not be discussed here again. TCNB was mainly chosen as a quencher, due to the lack of strong absorption of its reduced species in the spectral region beyond 460 nm. This allows for unperturbed investigation of the expected bands for twofold oxidized OTA at 600 nm and especially in the NIR region. The traces depicted on the left of Figure III.33 correspond to the transient absorption spectra of Ru–OTA in presence of an excess of TCNB. Excitation occurred at 532 nm with a nanosecond pump-probe setup. Hereby, the blue spectrum was recorded at a low excitation power of 10 mJ, whereas the cyan one was recorded at a laser power of 75 mJ. Both spectra show similar absorption features. The rather strong absorption from 378 to 459 nm is indicative of TCNB – formation.116 Furthermore, the transient spectrum reveals a broad band stretching over the whole NIR region with a local maximum at 1300 nm. This feature can be attributed to singly oxidized OTA, usually also exhibiting a band at 430 nm which is superimposed by the TCNB – band in the present case. Between 500 and 800 nm a more or less steady rise of the absorption is observed. However, no clear bands can be distinguished. In general, no distinct differences between the low- and high-power transient spectrum are observable, apart from the difference in intensity. As a result, the most likely photoproduct in the present flash-quench experiment is Ru–OTA+. On the right side of Figure III.33 the decays of the absorption signals at 378, 459, and 795 nm are depicted. The first two correspond to signals of reduced TCNB and the one at 795 nm

72

normalized optical density

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

100

mDOD

50 0 -50 -100 400

600

800

1000

1200

wavelength / nm

1400

1600

1

378 nm 459 nm 795 nm

0.1

0.01 0

50

100

150

200

250

300

time / µs

Figure III.33 Left: Power-dependent transient absorption spectra of Ru–OTA in presence of an excess of TCNB after excitation at 532 nm. Blue: 10 mJ; cyan: 75 mJ. Right: Decays of the transient absorption signals at 378 (blue), 459 (cyan), and 795 nm (orange). All spectra were recorded on de-oxygenated MeCN solutions.

is assigned to OTA+, as discussed above. As a result, the signals should exhibit the same lifetime. Fitting of the decays gives two different lifetimes, 13 µs and 67 µs. The need for double-exponential fitting can be explained by the presence of trace amounts of free [Ru(bpy)3](PF6)2 which also undergoes flash-quenching with TCNB. Therefore, the short-living species is assigned to Ru–OTA+, whereas the longer-living species is associated with [Ru(bpy)3]3+ which has a lifetime of 64 µs in flash-quenching experiments with TCNB. In order to determine, if hole accumulation on the OTA unit can indeed be excluded the difference spectrum between the high and the low-energy spectrum is constructed. Prior to performing the subtraction, the spectra have to be normalized to a point at which they show the same photo-generated species. In the present case, the TCNB – band at 459 nm in the visible region of the transient spectrum was used for this purpose. Due to absorption of both OTA+ and OTA2+ over the whole spectral range, the NIR region was normalized to the isosbestic point (1285 nm) between the two differently oxidized amine species. Multiplication of the low-energy spectrum by 7.5 (quotient of 75 mJ and 10 mJ) produced a modified low-energy spectrum with a higher overall optical density than the spectrum at 75 mJ exhibited. This behavior is most likely due to saturation effects at such high energies. The sharp band at 1064 nm is due to scattered light from the laser beam which comes to effect because of the poor overall signal-to-noise ratio in the difference spectrum. However, three distinct bands at 380, 555, and 1050 nm can be observed in the difference spectrum. Compared to the difference spectra as obtained by chemical oxidation of OTA to OTA+ (grey trace in Figure III.34) and OTA2+ (black trace in Figure III.34), the transient spectrum shows similarities. The orange spectrum does show increased absorption between 500 and 600 nm which might indicate the presence of twofold oxidized OTA. This assumption would be supported by the band peaking at ≈ 1060-

73

mDDOD

DOD

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.6 0.4 0.2 0 -0.2 6 4 2 0 -2 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.34 Top: Difference Spectra of OTA+ (grey trace) and OTA2+ (black trace) obtained by chemical oxidation with Cu(ClO4)2. Bottom: Difference spectrum of the high- and low-energy spectra as shown in Figure III.33 after normalization.

1070 nm in the NIR region, although the expected band maximum for OTA2+ is at 1130 nm. Measurements in the NIR region are performed at 10 nm intervals with a set bandwidth of 10 nm to ensure a measurable signal. Additionally, if twofold oxidized OTA were present in the here performed measurement, the transient decays would be expected to be triple exponential. As discussed above, the longest living species (67 µs) is due to flash quenching of [Ru(bpy)3]2+ with TCNB and the main photoproduct Ru–OTA+ displays a lifetime of 13 µs. In flash-quench experiments diffusion always plays a pivotal role and is responsible for the prolongation of the charge-separated state lifetime. Possible formation and back electron transfer reaction of the desired second charge-separated state, Ru– OTA2+ is also diffusion controlled in this case and might therefore exhibit a similar lifetime than the first charge-separated state. If Ru–OTA2+ is indeed formed in the present flash-quench experiment it will be present to a much smaller amount than the first charge-separated state, due to the multitude of deactivation processes that are open to it. The smaller percentage would explain the poor signal of the difference spectrum, as well as the lack of a third species in the kinetic absorption. However, exclusive assignment to Ru–OTA2+ is still questionable and will have to be further investigated by pump-pump-probe experiments.

III.4.6 Time-Resolved Study of AQ–Ru–OTA in Pure MeCN Regarding the thermodynamics, charge accumulation in AQ–Ru–OTA in pure MeCN solution is deemed rather unlikely (Figure III.29). Especially, considering the large negative potential for the second reduction of AQ. Figure III.35 depicts the data obtained for the time-resolved transient experiments on the triad in pure, de-oxygenated MeCN (10−5 M). Excitation occurred at 532 nm and the spectra were obtained at excitation powers of 2.1 (blue) and 58 mJ (cyan). The cyan transient spectrum exhibits a

74

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

150 100 50 0 -50 -100 30 20 10 0 -10 -20 -30

1

normalized intensity a.u.

mDDOD

mDOD

bleach at 340 nm due to bleaching of ground state absorptions of AQ and OTA observable in the UV-Vis spectrum in Figure III.27. The band at 415 nm can be assigned to singly oxidized OTA, as can be derived from the difference spectra after chemical oxidation of OTA with Cu2+. This observation is supported by the constant rise in optical density from 630 nm on to longer wavelengths. In combination with the broad band stretching over the whole NIR region, also indicative of OTA+ and the band at 565 nm attributed to AQ –,113 the observed photoproduct is AQ––Ru–OTA+. All transient absorption decays on the right side of Figure III.35 can be fitted single-exponentially, resulting in a lifetime of 2.4 µs for the first charge-separated state. 415 nm 565 nm 685 nm

0.1

0.01 400

600

800 1000 1200 wavelength / nm

1400

1600

0

2

4

6 time / µs

8

10

12

Figure III.35 Left: Transient absorption spectra after excitation of a 10−5 M MeCN solution of AQ–Ru–OTA at 532 nm. High and low-energy spectra are shown with cyan (58 mJ) and blue (2.1 mJ) traces. Difference spectrum after normalization at the isosbestic points and subtraction of the blue from the cyan spectrum (orange trace). Right: Transient absorption decays of the signals at 415, 565, and 685 nm.

For the construction of the difference spectrum (orange trace), the high and lowenergy spectra were normalized to the band at 565 nm or to the isosbestic point at 1285 nm. Though the lifetime of AQ––Ru–OTA+ is long enough to allow for a second excitation within the same 10 ns laser pulse, formation of CSS2 is unlikely. In the present system in pure MeCN, AQ cannot be reduced a second time by the reduced or excited ruthenium photosensitizer. As a result, the sole photoproduct is indeed AQ–– Ru–OTA+ as can be derived from the difference spectrum. No spectral indications of OTA2+, such as increased absorption at 600 or 1130 nm are observable.

III.4.7 Time-Resolved Study of AQ–Ru–OTA in Presence of L EWIS Acids In the AQ–Ru–OTA triad, the addition of acids is necessary for shifting the second reduction potential of AQ to more positive potentials. Only then can the excited or reduced ruthenium sensitizer forward another electron to AQ –. In Section III.2.4 the Brønsted acid-induced spectral changes in the OTA absorption was discussed. Despite the expectation, strong acids such as triflic acid seemingly promoted oxi-

75

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes dation of the donor. However, only chloroacetic acid was used in the time-resolved experiments investigating AQ–Re–OTA. Though, no premature oxidation of OTA was observed with chloroacetic acid, no twofold charge separation was achieved. By using additives that would bind stronger and more exclusively to AQ, in combination with higher excitation densities, twofold charge separation might be achievable. Instead of Brønsted acids, Lewis acids were chosen as additives. Rare earth metal ions, should bind solely to AQ and not influence OTA. Metal Ion-Promoted Electron Transfer (MCET)

Fukuzumi et al. have performed extensive studies on the mechanics of the so-called metal ion-coupled electron transfer and its influence on photoinduced electron transfer reactions.133–135 With the help of Lewis acids, electron transfer can be initiated that would otherwise not occur, so-called metal ion-induced electron transfer. Photoinduced electron transfer from C60 to p-chloranil (Cl4Q) is endergonic and does not occur in pure benzonitrile. However, in presence of Sc3+ the reduction potential of Cl4Q is less negative which enables the electron transfer reaction (Scheme III.13 top).134 O

O Cl 3C * 60

Cl

+ Cl

+ Sc3+

C60

Cl

Cl

Cl

Cl

Cl O Sc3+

+

O hν

Cl4Q

C60 O + O Fe

O O

N H

Fe

N H

+ Sc3+

(CH2)5 O

Fc-NQ

O

O

O Sc3+

Scheme III.13 Top: Photoinduced MCET in a bimolecular system of C60 and Cl4Q after addition of a Lewis acid. Bottom: Thermally induced MCET in a Fc–NQ dyad after addition of a strong Lewis acid.

Not only can photoinduced transfer of an electron be rendered possible by addition of metal ions, they can also induce thermal electron transfer in otherwise inert systems (Scheme III.13 bottom). In a ferrocene-naphthoquinone dyad, Fc–NQ, no thermal electron transfer occurs at RT in MeCN. In presence of Mg(OTf)3, still no charge transfer product could be observed. Only upon addition of Sc3+ oxidized ferrocene and reduced NQ was formed. The NQ radical anion was stabilized by complexation with the metal ion.136 Consequently, the ability to induce or promote charge transfer is

76

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes strongly dependent on the Lewis acidity of the metal ion. This is very similar to what is observed for Brønsted acids. Similar to PCET, MCET can occur sequentially starting with electron transfer (ET), followed by metal binding (MB) to the reduced acceptor. A concerted mechanism is also possible in which electron transfer and metal binding occur simultaneously.133 Usually, no binding of the metal ion to the neutral acceptor is observed (Scheme III.14).

D*-A +

D -A

n+

+M +

-

DEred n+

D -A --M

n+

-M

D-A

Scheme III.14

Energy scheme for metal-ion coupled electron transfer after photoexcitation yielding D+–A–-Mn+. Metal ion-decoupled back electron transfer produces the dyad in its ground state.

Binding of the metal ion leads to a decrease in driving force for the back electron transfer (∆Ered ) due to their strong electrostatic interaction with the reduced acceptor. Setting the dyad back to its ground state requires the removal of the electrophilic metal ion from the reduced acceptor, which is referred to as metal ion-decoupled electron transfer (MDET).134 Transient Photoproducts in Presence of Different L EWIS Acids

As discussed above, Lewis acids of different strengths might influence the nature of the charge-separated state in AQ–Ru–OTA and may even promote formation of the CSS2 without premature oxidation of OTA. For the experiments metal ions of various Lewis acid strengths were chosen. Their Lewis acidity can be derived from their charge and ion radius.135 The higher the charge and the smaller the metal ion, the stronger a Lewis acid and the stronger it binds to negative charge. All used metal ions are listed in Table III.4 in descending order of strength, together with their ion radii. Table III.4 Used metal ions and their radii.137

metal ion

Sc3+ Eu3+ Mg2+ Ca2+ Ba2+

radius / Å

0.81

1.07

0.89

1.12

1.42

77

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes Figure III.36 displays the transient absorption spectra recorded on de-oxygenated solutions of the triad in MeCN. Complex concentrations were 2.5 · 10−5 M and of each Lewis acid an excess of 400 eq was added. Recording of the spectra occurred in a 200 ns time window directly after excitation with a 532 nm laser pulse. 100

AQ-Ru-OTA +Ba2+ +Ca2+ +Mg2+

80

mDOD

60 x5

40

x5

20 0

+Eu3+ +Sc3+

-20 -40 -60 300

400

500 600 wavelength / nm

700

800

Figure III.36 Transient absorption spectra of de-oxygenated 2.5 · 10−5 M MeCN solutions of AQ–Ru–OTA with 400 eq of Sc3+, Eu3+, Mg2+, Ca2+, and Ba2+. For better visualization the red and green spectra were multiplied by a factor of 5. The spectrum of the pure triad is shown as reference (black).

All spectra recorded in presence of the metal ions do in principle display the same spectral features, a bleach at 335 nm and bands at 420 nm and around 550 nm, as the triad in pure MeCN. While the bleach at 335 nm can again be attributed to ground state bleaching of OTA and AQ, the presence of the strong band at 420 nm and the lack of increased absorption around 600 nm does indicate the presence of OTA+ in the photoproduct. Furthermore, the still present band around 550 nm, indicative of AQ –,113 supports the assumption of the first charge-separated as the main photoproduct. The weaker Lewis acids Mg2+, Ca2+, and Ba2+, induce a slight blue-shift of the AQ – band over 20 nm, which might be attributable to a very weak coordination of the metal ions to the reduced AQ. Hydrogen bonding between AQ – and e.g. H2O, however, induces a more pronounced hypsochromic shift of the band to 500 nm.92 The stronger Lewis acidic ions Sc3+ and Eu3+ significantly reduced the optical density of the transient spectrum and are multiplied by a factor of 5 for better display and to enable band assignment. Again, a bleach at 330 nm and two bands at 410 and 570 nm are observed. Though, the bands are shifted compared to the spectra with the weaker coordinating metal ions, no spectral indication for twofold oxidized OTA is observable. As a result, the predominant photoproduct in all cases is most likely the first charge-separated state AQ––Ru–OTA+. The reason for the missing shift for the band assigned to AQ – in presence of the strong Lewis acids is unclear. If already the weaker binding metal ions induce a slight shift, the effect would be expected to be even more severe for the stronger ones, similar to the observations in III.2.5 in presence of different Brønsted acids. Another

78

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes interesting aspect is the decrease in optical density stemming solely from the Lewis acidity of the metal ions. Neither the solvent, nor the concentrations were changed in the measurements. If the lifetime of the photoproduct is very short the optical density of the transient signal will be very low because detection with the nanosecond setup will be difficult. On the other hand, an extremely long lifetime of the photoproduct, exceeding the detection limit of the instrument setup will also result in a weak signal. At some point, no more photoproduct can be produced to a sufficient amount in the transient experiment, due to to its very slow decay to the ground state. Therefore, not enough triad molecules in the ground state are left before the next excitation pulse. Kinetic Absorption Behavior in Presence of Different L EWIS Acids

In the following Figure III.37 the decays of the transient absorption signals in the above discussed spectra are depicted. Compared to pure MeCN solution (2.4 µs), AQ––Ru–OTA+ exhibits a slightly increased lifetime of 3.0 µs in presence of Ba2+ (a) or Ca2+ (b), when only the signals around 420 and 690 nm are considered. They show a clear single-exponential decay. Thus, the added metal ions do stabilize CSS1 to a certain extent. For the AQ related bands at 560 and 565 nm, however, a clear deviation from the single-exponential character of the decays is observed. Fitting of the decays gives lifetimes of 139 ns for an additional short-living species in presence of Ca2+ and 223 ns in the Ba2+ containing sample. In the present case no band shift was observed for the AQ-related bands in presence of the Lewis acids (Figure III.36), as was observed for AQ–Re–OTA in presence of Brønsted acids. Coordination of the metal ions might not induce a hypsochromic shift of the AQ– absorption band. If the band for AQ– and metal-coordinated AQ– are superimposed, this might explain the bi-exponential decays only present for the AQ–-related band around 560 nm. In presence of Mg2+, also a weaker Lewis acid, the same deviant behavior is observed for the AQ related band and the lifetime of the charge-separated state is only 2.6 µs. Clear bi-exponential character is observed and the fitted lifetimes are 172 and 392 ns, respectively. These, in combination with the very weak transient signal, do not meet the expectations of stabilization of the charge-separated state and might be due to metal ion-promoted sample degradation. However, addition of Sc3+ did lead to a significant increase in lifetime as can be seen in Figure III.37 e. After photoexcitation, two species are formed, one of which does not decay within the detection range of the nanosecond setup. Double-exponential fitting gives lifetimes of ≈1 and 20 µs for the two species. The real lifetime of the long-living species might indeed be different, due to a high uncertainty in this fit. However, the lifetime of the shorter living species resembles the one of AQ––Ru–OTA+ in pure MeCN. The longer decay compound

79

normalized optical density

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 1

1

420 nm 560 nm 690 nm

0.1

0.01

1

425 nm 565 nm 683 nm

0.1

0.1

a)

b)

0.01

5

10 time / µs

normalized optical density

0

15

20

413 nm 560 nm 685 nm

c)

0.01

5

0

1

10 time / µs d)

15

5

0

20

1

10 time / µs

15

20

e)

0.1

413 nm 565 nm 770 nm

0.01 0

413 nm

0.1 1

2

3

0

4

10

20

time / µs

30

40

50

time / µs

Figure III.37 Decays of the transient photoproduct in presence of various metal ions. a) Ba2+, b) Ca2+, c) Mg2+, d) Eu3+, and e) Sc3+.

might indicate the presence of a charge-separated state which is strongly stabilized by tight binding of Sc3+ to the reduced AQ acceptor. Sc3+ Concentration-Dependent Time-Resolved Experiments

1

0.5eq 1eq 4eq 40eq 400eq

0.1

1.0 0.8

kET / s-1

normalized optical density

Due to the unusually long lifetime of the charge-separated state AQ––Ru–OTA in presence of an excess of Sc3+ the influence of this metal ion on the transient photoproduct was investigated further. Concentration-dependent experiments were performed in de-oxygenated MeCN. On the left of Figure III.38 the emission decays of the triad in presence of varying concentrations of Sc3+ are depicted.

0.6 0.4 0.2 0

0.01 0

200 400 600 800 1000 1200 1400 1600 1800 2000 time / ns

0

2

4 6 8 c(Sc3+) / mM

10

Figure III.38 Emission decays of AQ–Ru–OTA with increasing Sc3+ concentrations. Right: Plot of kET vs c(Sc3+).

The origin of the second, long-lived luminescent species can be attributed to highly

80

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes luminescent impurities, as discussed in Figure III.30. Therefore, the short lifetimes were used for determination of the electron transfer rate constants. The latter is plotted vs the Sc3+ concentration. Linear fitting gives a basically non-existent slope and the electron transfer rate is therefore not dependent on the Sc3+ concentration. This further indicates that metal ion binding occurs after electron transfer, the rate determining step.138 Addition of scandium ions is not necessary to initiate electron transfer in AQ–Ru–OTA since charge separation is already observed in pure MeCN. Thus, the fit on the right of Figure III.38 does not intercept at the origin. The overall spectral shape for the scandium ion concentration-dependent measurements is indicative of AQ––Ru–OTA+ formation as the main photoproduct due to the OTA+ band at 420 nm and the AQ–-related band at 565 nm. However, with increasing amounts of Sc3+ the optical density decreases in the transient absorption spectra as depicted on the left of Figure III.39. 2.5

100

mDOD

0

0.5eq 1eq 4eq 40eq 400eq

-50 300

400

500 600 wavelength / nm

700

800

absorbance

2.0 50

1.5 1.0 0.5 0 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.39 Left: Transient absorption spectra after excitation of AQ–Ru–OTA at 532 nm in de-oxygenated MeCN with increasing amounts of Sc3+ (0.5400 eq). Right: Ground state absorption spectra of the sample solutions. The color code is the same as on the left and the spectrum in pure MeCN is depicted in black for comparison.

It was further noticed that addition of increasing amounts of Sc3+ to the complex solutions seems to initiate premature oxidation, prior to the transient experiments, as depicted on the right side of Figure III.39. In presence of 40 eq Sc(OTf)3 (red trace) ≈ 77% OTA+ had already been formed. The percentage was calculated in relation to the amount of OTA+ formed in the chemical oxidation experiment. The decrease in optical density correlates rather well with the increasing OTA+ formation. Scandium ions show a very high affinity towards oxygen and facilitate the one-electron reduction of O2 and the sample solutions are mixed under ambient conditions prior to de-oxygenation. Already at this early stage thermal electron transfer from OTA to O2 might be induced, which may be responsible for the premature oxidation of OTA.136 The possibility of this assumption is further investigated in the continuous experiments discussed in III.6.

81

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

normalized optical density

Despite the unchanged transient absorption spectra, increasing amounts of Sc3+ induced prolongation of the charge-separated state lifetime. 1

0.1

0.01

pure 0.5eq 1eq 4eq 40eq 400eq 0

5

10 time / µs

15

20

Figure III.40 Kinetic absorption decays of the band at 420 nm (Figure III.39) in presence of various Sc3+ concentrations in de-oxygenated MeCN solution.

The transient absorption decays of the band at 420 nm depicted in Figure III.40 clearly show that addition of up to one equivalent of Sc(OTf)3 does not lead to significant changes in the charge-separated state lifetime. The obtained lifetime of 2.6 µs is comparable to the one obtained in pure MeCN (2.4 µs). Upon further increase of the scandium ion concentration, the decay character becomes increasingly double-exponential and significantly longer-living species are formed. Their lifetime increases with increasing scandium concentration to 60 µs in presence of a 40-fold excess. In presence of a 400-fold excess of Sc3+ the decay can not be fitted reliably anymore. However, we can conclude that the lifetime in this particular sample lies well in the microsecond, if not millisecond regime, since its complete decay cannot be detected with the used nanosecond setup. The presence of this very long-living photoproduct in the de-oxygenated AQ–Ru– OTA solution in presence of a large excess of Sc(OTf)3 lead us to perform continuous irradiation experiments. With such a long lifetime, photo-accumulation and tracking of the photoproduct on a UV-Vis timescale should be possible. Regarding the premise of premature oxidation of OTA to a certain extent during sample preparation, reducing agents were added. Alas, neither the addition of cobaltocene, nor Na/Hg lead to the desired re-reduction of OTA. The amount of prematurely oxidized OTA is dependent on the scandium concentration and so is the lifetime of the photoproduct. Therefore, finding a balance between the amount of OTA+ present prior to the experiments and a sufficiently long lifetime of the generated photoproduct was crucial. The exact parameters for the continuous irradiation experiments and their results are discussed in III.5 and III.6.

82

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.4.8 Time-Resolved Studies on AQ–Ru–OTA in presence of Sc(hmpa)33+ For all time-resolved transient experiments above, Sc(OTf)3 was used as the metal ion source. When the triflate salt is dissolved in MeCN the OTf– ligands are abstracted and the metal ion exists in its “naked” form as Sc3+. Metal ions display a very high affinity towards oxygen and scandium seemingly induced premature OTA oxidation, especially in this naked form, as discussed above. Since OTA+ formation in the scandium containing samples proceeded much faster and in higher yield, even the added reductants, cobaltocene or Na/Hg, were not able to counteract and re-reduce OTA+. Therefore, the Lewis acidity of free Sc3+ was attempted to be masked by attachment of stronger-binding ligands, such as hexamethylphosphoramide (hmpa). Hmpa as a rather bulky, monodentate ligand limits the coordination number to six for scandium ions and the resulting complex is very stable regarding ligand loss.139 Preparation of Sc(hmpa)33+ occurred by heating a THF solution of Sc(OTf)3 and 3 eq hmpa to 50 °C under N2 atmosphere. Sc(hmpa)33+ precipitates from the cooled reaction mixture in form of the triflate salt. Separate de-oxygenation of the scandium and triad solution and the use of only 20 eq of Sc(hmpa)33+ resulted in a decrease of premature OTA+ formation to 38%. However, complete prevention of premature oxidation was not possible. In Figure III.41 the transient low- (10 mJ, blue) and high-energy (80 mJ, cyan) spectra are depicted on the left together with the difference spectrum. The latter was obtained after normalization of the high and low energy spectra to the isosbestic points between the OTA+ and OTA2+ species as observed in the chemical oxidation experiment (III.4.4). Due to OTA+ absorption in the whole spectral range the isosbestic points of OTA+ and OTA2+ seemed to be the most logical choice for normalization. normalized optical density

mDOD

200 100 0

mDDOD

-100 8 4 0 -4 400

600

800

1000

1200

wavelength / nm

1400

1600

1

415 nm 565 nm 770 nm

0.1

0.01 0

5

10 time / µs

15

20

Figure III.41 Left: Transient absorption spectra at high (80 mJ, cyan) and low (10 mJ, blue) pulse energies after excitation of de-oxygenated MeCN solutions at 532 nm. Difference spectrum after normalization at 478 and 1285 nm and subtraction of the low from the high energy spectrum. Right: Kinetic absorption decays of the photoproduct at different wavelengths.

83

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes The spectra recorded at high and low energies display the same transient absorption features which indicates the presence of the same photoproduct in both cases. Bands at 415 and 1300 nm indicate once again the formation of OTA+. Together with the band at 565 nm assigned to the AQ radical anion, the main photoproduct in presence of Sc(hmpa)33+ is most likely AQ––Ru–OTA+. Coordination of the scandium complex to the reduced AQ should result in an increase of the charge-separated state lifetime, as was observed for the use of different metals. Fitting of the decays, however, gave a mean lifetime of 2.1 µs for the photoproduct which is the same as for the triad in pure, de-oxygenated MeCN (2.4 µs), within the accuracy of the measurement and fitting procedure. Even in presence of the weaker Lewis acids Ba2+ and Ca2+ the lifetime of AQ––Ru–OTA+ was increased to 3.0 µs. It is therefore concluded that the scandium-hmpa complex is not coordinated to the reduced AQ in the photoproduct. The difference spectrum (black trace in Figure III.41 displays bands at 400, 550, and 1000 nm. Twofold reduced AQ shows absorption at 470 and 622 nm in DMF,90 however neither one of the observed bands in the difference spectrum would fit to this assignment. Furthermore, the band for OTA2+ is expected at 1130 nm in the NIR region together with increased absorption around 600 nm. Though the observed band at 550 nm might be interpreted in terms of twofold oxidized OTA, the band at 1000 nm shows too large a shift in relation to the expected maximum at 1130 nm.

III.4.9 Time-Resolved Studies in Presence of N,N-Diphenylthiourea In earlier attempts to stabilize the first charge-separated state in order to promote formation of the second one by protons, hydrogen-bonds, or rare earth metal ions (Lewis acids) the additives seemed to induce premature oxidation of the OTA donor. Therefore, an additive was chosen that should only interact with the AQ acceptor. N,N-diphenylhiourea is known to recognize benzoquinones, such as ubiquinone 0 (Q0) and form a stable complex comprised of two diphenylthiourea molecules and a twofold reduced ubiquinone (Q02–) as shown in Scheme III.15.140 Compared to dialkylthiourea, diphenylthiourea is more acidic and a better hydrogen bond donor. In electrochemical measurements no formation of the twofold reduced ubiquinone was observed in presence of an alkylthiourea derivative. Additionally, the chelating effect proved to be essential for the twofold reduction. Therefore, diphenylthiourea, was added to the triad solution. Addition of N,N-diphenylthiourea should not result in oxidized OTA prior to the laser experiments and it should bind exclusively to reduced AQ, facilitating its second reduction and also stabilizing the final charge-separated state.

84

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes S N H O

O H3CO

e

e

H3CO H3CO

O

O H N

Q0

N H O

H3CO H3CO

H3CO

Scheme III.15

S N H

N H

O H N

H N

H N

S

S

Q0

Q02-

Stepwise reduction of ubiquinone 0 (Q0) and coordination with N,Ndiphenylthiourea.

Figure III.42 depicts the ground state absorption spectra of the triad in pure MeCN (black trace) and in presence of 1000 eq diphenylthiourea (grey trace). Due to the excess of the thiourea derivative absorption is increasing rapidly below 350 nm. extinction coefficient / 103 M-1cm-1

100 80 60 40 20 0 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.42 Ground-state absorption spectra of AQ–Ru–OTA in pure MeCN (black trace) and in presence of 1000 eq of diphenylthiourea (grey trace).

The absorption spectra clearly show no increased absorption around 1300 nm due to premature oxidation of the OTA donor. Therefore, if bands indicate the presence of twofold oxidized OTA in the time-resolved experiments, it can be assumed that its formation is indeed a two-photon process. In the time-resolved absorption experiments the photoproduct as displayed in Figure III.43 is formed within the pulse duration of the laser (10 ns). Both, the high and low energy spectrum display the similar absorption features in the visible region. The band at 418 nm together with the steady rise of optical density beyond 600 nm are clear signs for OTA+ (Figure III.31). As counterpart to singly oxidized OTA the singly reduced AQ is expected. The band at 538 nm does not fully coincide with the usual AQ – band shift of 565 nm or the protonated version, AQH, around 400 nm.90,93 For hydrogen-bonded, singly reduced AQ a band at 500 nm is expected.92 Nevertheless, the AQ related band is shifted slightly towards shorter wavelengths settling in between the the singly reduced and hydrogen-bonded state. Therefore, the reduced AQ acceptor seems to be only weakly coordinated to diphenylthiourea. In the NIR region

85

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

mDDOD

mDOD

the spectra recorded at high and low energies clearly show differences in their signature. At low energy (12 mJ) the very broad band with a local maximum at 1310 nm is visible which is typical for singly oxidized OTA+ as determined by chemical oxidation of the donor. The same experiment at a higher excitation density of 66 mJ changes the spectrum and the maximum shows a hypsochromic shift to 1190 nm. OTA2+ is known to show an absorption maximum at 1130 nm (Figure III.31). Compared to the band maximum of the transient experiment this gives a deviation of 60 nm. The NIR detector used to record the spectra between 800 and 1700 nm does so in 10 nm intervals. Furthermore, the bandwidth is set to 10 nm to allow for a recordable signal. Due to these necessary settings the NIR detector displays a rather low sensitivity which does lead to a certain insecurity in the assignment of the band maxima. However, a deviation of 60 nm is too large to be explained by the uncertainty of the detector. 40 30 20 10 0 -10 10 5 0 -5 -10 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.43 Transient absorption spectra of AQ–Ru–OTA in presence of 1000 eq diphenylthiourea after excitation at 532 nm. Top: high energy spectrum = cyan (56-66 mJ), low energy spectrum = blue (11-12 mJ). Bottom: Difference spectrum obtained by subtraction of the low from the high energy spectrum after normalization to the isosbestic points at 478 and 1285 nm.

After subtraction of the spectral features of AQ––Ru–OTA+ only bands assignable to the desired second charge-separated state AQ2––Ru–OTA2+ should be present. Therefore, the high and low energy spectra were normalized to the isosbestic points between the OTA+ and OTA2+ species, and the low-energy spectrum was subtracted from the high-energy spectrum. As discussed above, the former represents the first charge-separated state. The bottom part of Figure III.43 displays the difference spectrum created in this manner as an orange trace. It only displays two bands at 420 nm and a rather broad band in the NIR from 800 to 1300 nm. Beyond 1300 nm negative optical density is observed. This might indicate formation of OTA+ and subsequent further oxidation to OTA2+ which results in a decrease of OTA+ corresponding absorption beyond 1300 nm. If the band in the NIR is considered a sign for OTA2+, another band around 600 nm would be expected. However, the difference spectrum does not show increased absorption in said spectral range. Furthermore, the approximate maximum of the broad band is at 1050 nm which is off by 80 nm compared to

86

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

normalized optical density

the expected wavelength of 1130 nm. However, some deviation in the band maxima might again be due to the detector settings required for the NIR measurements. From the transient absorption experiments at hand it is therefore concluded that the only photoproduct generated is the singly charge-separated state AQ––Ru–OTA+. If AQ2––Ru–OTA2+ is indeed formed in the measurement with a high excitation density, it should also be visible in the transient absorption decays. Here, a bi-exponential decay indicating the two different species is expected. Figure III.44 depicts the decays of the bands from the transient absorption experiment. 1

418nm 538nm 680nm 780nm

0.1

0.01 0

2

4

6 8 time / µs

10

12

14

Figure III.44 Absorption decays of the bands at 418, 538, 680, and 780 nm recorded after excitation of the de-oxygenated, diphenylthiourea containing MeCN solution with a 532 nm laser pulse (56 mJ).

All bands show unmistakable single-exponential decay behavior after the excitation with a high-energy laser pulse. The lifetime is given as 2.5 µs which is very similar to what was obtained for the triad in pure MeCN solution (2.4 µs). From the decay behavior, in conjunction with the transient absorption spectra, especially the constructed difference spectrum it can be concluded that even at high excitation powers the predominant photoproduct is CSS1 AQ––Ru–OTA+. In addition, the very similar lifetimes of the first charge-separated state with and without an excess of diphenylthiourea indicates only very slight coordination of the thiourea to the reduced AQ acceptor. This was already proposed above due to the shift of the AQ– related band to 538 nm.

III.4.10 Summary on Electron Transfer in AQ–Ru–OTA In the preceding section electron transfer in a AQ–Ru–OTA triad was studied. Some problems encountered in the analog rhenium triad could be eradicated in AQ–Ru– OTA. From the absorption spectrum of the triad in MeCN, the 1MLCT band is now clearly discernible and not superimposed by OTA or AQ absorptions. This allowed for selective excitation of the ruthenium photosensitizer. Furthermore, the excitation wavelength of 532 nm used for ruthenium excitation allowed much higher photon fluxes. The high-energy measurements were performed at laser powers of 50-80 mJ

87

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes which translates to roughly 9 to 14 photons per pulse per molecule. From a theoretical viewpoint twofold excitation of the same complex molecule within the same high-energy laser pulse should therefore be possible. Photoexcitation of a Ru–OTA reference dyad lead to separation of a single charge, and the lifetime of the charge-separated state could be significantly prolonged from 10 ns to 13 µs in a flash quench experiment with TCNB. Spectral indications at 380, 555, and 1050 nm for twofold oxidized OTA were observed in the difference spectrum after normalization and subtraction of the low- from the high-energy transient spectrum. In the near future, Ru–OTA will be further investigated in flash-quench experiments with TCNB with the help of a pump-pump-probe laser setup. By this, we hope to find more conclusive evidence for the presence of Ru–OTA2+. Photoinduced electron transfer in AQ–Ru–OTA in de-oxygenated MeCN leads to the expected CSS1, AQ––Ru–OTA+, with a lifetime of 2.4 µs. Compared to the rhenium triad in pure, de-oxygenated MeCN this represents a 12-fold increase which is another advantage of this ruthenium complex. To achieve stabilization of the negative charge on AQ in CSS1 different metal ions were employed. Upon photoexcitation of the ruthenium photosensitizer, metal ion-coupled electron transfer (MCET) was expected. In presence of Brønsted acids, premature oxidation of OTA prior to the transient experiments was observed for AQ–Re–OTA. This was not expected to occur when using metal ions. Addition of weaker Lewis acids such as Ca2+, Ba2+ or Mg2+ did lead to a slight prolongation of the charge-separated lifetime due to their coordination to the singly reduced AQ. By this stabilization of CSS1, the lifetime was increased from 2.4 µs in pure MeCN, to 3.0 µs. However, formation of CSS2 was not observed. Presence of the very strong Lewis acid Sc3+ lead to formation of a extremely long-living species that did not show complete decay within the experimental detection limit. Scandium concentration-dependent experiments revealed a decrease in optical density of the transient absorption signal with increasing scandium concentration. Furthermore, a concomitant increase in prematurely formed OTA+ was observed. In the case of AQ–Ru–OTA, electron transfer is the rate-determining step for the formation of the photoproduct, not binding of the metal ions to AQ–. This is verified by the independence of the electron transfer rate from the scandium concentration. The oxidation behavior in presence of an excess of Sc3+ is further investigated with continuous irradiation experiments in III.6. The problem regarding premature OTA+ formation due to the added scandium ions could not be solved by attaching the bulkier hmpa ligands to the metal ion. The lifetime of the charge-separated state AQ––Ru–OTA+ in presence of Sc(hmpa)33+ is 2.1 µs.

88

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes This indicates that in its masked form, the scandium ions do not exert any influence on the formed photoproduct. Only the use of N,N-diphenylthiourea did not lead to any spectral changes due to OTA oxidation. The thiourea derivative is known to show exclusive affinity to reduced quinones. In presence of a 1000-fold excess, the lifetime of the predominantly generated photoproduct, AQ––Ru–OTA+, is 2.5 µs. Seen in the single-exponential character of the absorption decays, only one species is present. Though no significant increase in lifetime was achieved by addition of diphenylthiourea, at high excitation densities of 66 mJ distinctive differences in the spectral signatures, especially in the NIR region are observed. Therefore, further photophysical investigation of AQ–Ru–OTA in de-oxygenated MeCN in presence of an excess of diphenylthiourea with a pumppump probe setup might give a more detailed insight. In this case the triad could be excited by to temporally delayed laser pulses that might even exhibit different excitation wavelengths. Use of two separate pulses should also allow for a clearer distinction of the one- and two-photon products.

89

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.5 Ferrioxalate Actinometry For all continuous irradiation experiments performed on solutions of AQ–Ru–OTA a Fluorolog-322 from Horiba Jobin-Yvon was used. Excitation occurred at 450 nm into the metal-to-ligand charge transfer band of the ruthenium photosensitizer. The entrance slits were opened to the maximum, a bandwidth of 14.7 nm, to allow as much light from the Xe-lamp as possible into the sample solution. In order to determine the exact photon flux and the number of photons available for the excitation of AQ–Ru–OTA, chemical actinometry was performed. A very convenient actinometer for wavelengths up to 450 nm is potassium ferrioxalate. It is characterized by its high quantum yield of photodegradation. Basically 100% of the absorbed photons lead to formation of Fe2+. The quantum yield of the degradation reaction steeply decreases at wavelengths higher than 450 nm. At this wavelength, however, the quantum yield is still 1.0.141,142 Therefore, ferrioxalate actinometry exhibits excellent applicability for the determination of the photon flux in a sample.

2 [Fe(C2O4)3]3–



2 Fe2+ + 5 C2O42– + 2 CO2

(III.3)

Upon irradiation of an acidic ferrioxalate solution, Fe3+ is reduced to Fe2+ and part of the oxalate ligand is oxidized to CO2. The simplified chemical equation for photoinduced ferrioxalate degradation is shown in Equation III.3.143 The photochemically produced Fe2+ subsequently forms an orange-red complex with 1,10-phenanthroline which can be quantified spectroscopically (λmax = 511 nm). Table III.5 Sample preparation for the calibration curve.

sample

1 2 3 4 5 6 7

(NH4)2Fe(SO4)2 phenanthroline / mL / mL 0 0.5 1 1.5 2 2.5 3

1 1 1 1 1 1 1

buffer / mL

H2SO4 / mL

H2O / mL

2.5 2.5 2.5 2.5 2.5 2.5 2.5

5 4.5 4 3.5 3 2.5 2

1.5 1.5 1.5 1.5 1.5 1.5 1.5

Preceding the actual actinometric measurement, a calibration curve needed to be

90

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes recorded to determine the extinction coefficient of [Fe(phen)3]2+ at 511 nm. For this, four solutions had to be prepared. A 4 · 10−4 M solution of (NH4)2Fe(SO4)2 in 0.05 M H2SO4, a 6.7 · 10−3 M aqueous 1,10-phenanthroline solution, a 0.6 M acetate buffer, and 0.05 M H2SO4. From these solutions and de-ionized H2O, six samples with varying Fe2+ concentrations were prepared according to Table III.5. Absorption spectra were measured of all samples and the absorbances at 511 nm were plotted vs the Fe2+ concentration. The slope of the fit in Figure III.45 gives the molar extinction coefficient of [Fe(phen)3]2+ at 511 nm to be 11500 M–1cm–1. 1.4

1.2

1.2

absorbance

1.0 0.8

increasing Fe concentration 2+

0.6 0.4 0.2 0

absorbance @ 511 nm

1.4

1.0 0.8 0.6 0.4 0.2 0

-0.2 400

450

500 550 600 wavelength / nm

650

700

0

20

40

60 80 100 -6 cFe2+ / 10 M

120

140

Figure III.45 Left: Absorption spectra of [Fe(phen)3]2+ with increasing concentrations. Right: Calibration points and fit for the determination of the molar extinction coefficient.

For the determination of the photon flux, a stock solution of 10 mL 1,10-phenanthroline solution (6.7 · 10−3 M), 25 mL acetate buffer (0.6 M), 40 mL diluted sulfuric acid (0.05 M), and 15 mL de-ionized H2O was prepared and 9 glass vials were filled with 9 mL each. The vials were covered with aluminum foil. In the dark, a 0.02 M potassium ferrioxalate solution in 0.05 M H2SO4 was prepared. Though, light-induced degradation of the ferrioxalate is key to the photon flux determination by actinometry, premature degradation is unwanted. Therefore, light exclusion during sample preparation is crucial. As a blank test, 1 mL of the non-irradiated ferrioxalate solution was added to a prepared glass vial and the absorption spectrum was recorded. Due to the lack of Fe2+ in the solution, no absorption of [Fe(phen)3]2+ is expected at 511 nm. Sample preparation for the irradiation measurements occurred as follows. 2 mL ferrioxalate solution were placed in a Quartz cuvette and irradiated for 10 s on the Fluorolog (λexc = 450 nm, slits = 14.7 nm). Subsequently, 1 mL was drawn from the cuvette, added to a prepared vial containing 9 mL of the above-mentioned phenanthroline-containing stock solution. After thorough mixing an absorption spectrum of the formed [Fe(phen)3]2+ was recorded. This procedure was repeated with irradiation times of 20, 30, 40, 50, 60, 90, and 120 s, respectively. The spectra obtained in this manner are depicted on the left in Figure III.46.

91

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 1.4

2.0

1.2

1.0

nFe2+ / 10-6 mol

absorbance

1.5

irradiation period

0.5

1.0 0.8 0.6 0.4 0.2 0

0 400

450

500 550 600 wavelength / nm

650

700

0

20

40

60 80 time / s

100

120

140

Figure III.46 Left: Absorption spectra recorded after increasing irradiation times. Right: Amount of Fe2+ (mol) vs the irradiation period.

With Equation III.4 the amount of Fe2+ produced by irradiation can be calculated. V1 is the irradiated volume, V2 the 1 mL sample drawn from the cuvette, A511 nm is the absorbance at 511 nm, d the pathlength (1 cm), V3 is the volume after complexation (10 mL), and  is the molar extinction coefficient (11500 M–1cm–1). n Fe2+ =

V1 · V2 · A511nm d · V3 · e511nm

(III.4)

By plotting the moles of Fe2+ vs the irradiation period the data points show a linear dependency. The photon flux can be derived from the slope of the fit, divided by the quantum yield (1.0). In the performed measurement this lead to a photon flux of photons 1.12 · 10−8 mols·hν , or 6.74 · 1015 s . This results in a total photon count of 5 photons per molecule per minute for the continuous irradiation measurements on AQ–Ru–OTA with the above given Fluorolog settings and the sample concentrations (10−5 M) and volumes (3 mL) used in the experiments described below.

92

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.6 Continuous Irradiation Experiments on Ru–OTA and AQ–Ru–OTA As discussed in III.4.7, strong Lewis acids such as Eu3+ or Sc3+ lead to the formation of a transient photoproduct with a very long lifetime, exceeding the detection limit of the nanosecond setup. By performing continuous irradiation (CI) experiments with a Xe-lamp as excitation source, we hoped to successively accumulate the long-living photoproduct and to follow its formation with UV-Vis absorption spectroscopy. All experiments on AQ–Ru–OTA were performed in aerated and de-oxygenated MeCN, in absence and presence of 0.02 M Sc(OTf)3, and under continuous irradiation and dark conditions. As a result, the dependence of the photoproduct on the O2 content, the Sc3+, and the light source could be evaluated. Furthermore, Ru–OTA was investigated in the same way to assess the influence of the AQ acceptor on the required irradiation times and the spectral differences. Sample Preparation Quick sample preparation and direct measurement proved to be crucial for all following experiments. Dry MeCN, stored over molecular sieve, was used as solvent. All aerated samples were prepared just prior to the measurements and used instantly. For the de-oxygenated samples the Sc(OTf)3, as well as the triad solution were de-oxygenated separately. This was achieved in a Schlenk-cuvette and mixing of both solutions occurred just before the measurement under inert conditions. The final concentrations were 10−5 M for AQ–Ru–OTA or Ru–OTA and 0.02 M Sc(OTf)3. This corresponds to an excess of 2000 eq Sc3+. Despite the fast preparation and separate de-oxygenation of the sample solutions, ≈ 30% OTA+ had already been formed prior to the start of the actual measurements. Residual O2 in De-Oxygenated Samples The Schlenk-cuvette used for the following measurements has a volume of 47 mL. With the help of the ideal gas equation p · V = n · R · T the moles of gas in the cuvette under ambient conditions can be calculated, which gives an amount of 1.91 · 10−3 mol; T = 298.15 K, R = 8.314 molJ ·K , and p = 100000 mN2 . Calculation of the residual O2 content was performed under the assumption that de-oxygenation of the sample solution by three freeze-pump-thaw cycles leads to a complete exchange of the ambient atmosphere with N2. The used N2 (PanGas 5.0) has an O2 content of ≤ 2 ppm, which means that 0.0002% of the gas in the cuvette is O2. This amounts to 3.8 · 10−9 mol residual O2 in the de-oxygenated sample solutions and an error of one

93

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes order of magnitude is assumed. This value is of interest for later interpretation of some experimental results. All spectra were recorded as difference spectra, subtracting the initial absorptions of the dyad and triad. If not state otherwise, the excitation wavelength was set to 450 nm. To determine the amounts of OTA+ and OTA2+ generated in the following experiments, the extinction coefficients of the chemical oxidation experiment were used. These are 17300 M–1cm–1 for OTA+ at 1310 nm, 59500 M–1cm–1 for the twofold oxidized OTA at 1130 nm, and each value corresponds to quantitative formation of the respective species. The color of the samples changed from a slight yellow for the non-oxidized species, to greenish-yellow in presence of OTA+, and greyish for twofold oxidized OTA.

III.6.1 Reference Experiments Figure III.47 depicts the obtained difference spectra of Ru–OTA (top) and AQ–Ru– OTA (bottom) after five successive UV-Vis measurements from 1700 to 200 nm. They were recorded immediately one after another. Each measurement takes approximately 3 min which results in a total irradiation time of 15 min. During this time the complexes are slightly irradiated by the light source of the UV-Vis spectrometer. Especially irradiation at 450 nm in the 1 MLCT band of the ruthenium photosensitizer might be problematic. Therefore, the spectral changes due to the UV-Vis measurement were evaluated prior to the actual CI experiments. Three general insights can be derived from the spectra. As stated above, during the UV-Vis measurements the complexes are irradiated by the light source of the spectrometer and this can lead to photochemistry. The spectra in Figure III.47 indicate the formation of OTA+, which represents part of the first expected photoproduct upon irradiation of AQ–Ru–OTA. Furthermore, the amount of formed OTA+ is higher in the experiments on the triad (Figure III.47 c and d). This is observed best in the spectra recorded on the de-oxygenated solutions, b and d in Figure III.47. On the basis of this, one can conclude that the attached AQ acceptor does influence the generation of the OTA+ photoproduct upon irradiation. During the same irradiation time, more than twice as much OTA+ was formed in AQ–Ru–OTA (18%), than in Ru–OTA (7%) in de-oxygenated solutions. The amount of OTA+ in the de-oxygenated solutions of AQ–Ru–OTA (5.4 · 10−9 mol) and Ru–OTA (2.1 · 10−9 mol) is similar to the amount of residual O2 (3.8 · 10−9 mol) in the sample solutions. In presence of O2 (Figure III.47 a and c) the difference in OTA+ formation between the dyad and triad is smaller. During the same irradiation time as above, 24% OTA+ is generated in Ru–OTA and

94

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.06

b)

a)

DOD

0.04 0.02 0 -0.02 -0.04 400 0.06

600

800 1000 1200 wavelength / nm

1400

1600

c)

400

600

800 1000 1200 wavelength / nm

1400

1600

400

600

800 1000 1200 wavelength / nm

1400

1600

d)

DOD

0.04 0.02 0 -0.02 -0.04 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.47 UV-Vis absorption spectra after five successive measurements. a) Ru– OTA + O2; b) Ru–OTA + N2; c) AQ–Ru–OTA + O2; d) AQ–Ru–OTA + N2. Complex concentrations were 10−5 M with 0.02 M Sc(OTf)3.

30% in the triad solution. This indicates that the OTA+ formation depends more on the amount of O2 present in the sample than on the AQ acceptor which leads to the conclusion that O2 is the terminal electron acceptor. To verify whether the mere presence of O2 is responsible for the oxidation of OTA or if OTA+ formation is indeed a light-induced reaction, another dark experiment was performed . For this, aerated solutions of Ru–OTA and AQ–Ru–OTA were prepared and a difference spectrum of each was recorded after 15 min in the dark (Figure III.48). The 15 min are the same duration as for the five successive UV-Vis experiments discussed above. 0.06

b)

a)

DOD

0.04 0.02 0 -0.02 -0.04 400

600

800 1000 1200 wavelength / nm

1400

1600

400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.48 UV-Vis absorption spectra of aerated MeCN solutions of a) Ru–OTA and b) AQ–Ru–OTA after 15 min in the dark (grey traces) and five successive measurements (black traces). Complex concentrations were 10−5 M with 0.02 M Sc(OTf)3.

The OTA+ contents received were 26% for the dyad and 32% for AQ–Ru–OTA. These

95

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes closely resemble the amounts of OTA+ after five successive UV-Vis measurements (black traces) and the deviation of 2% in both cases can be explained by the fact that new sample solutions had to be prepared. These reference experiments show that not the light source of the UV-Vis spectrometer, but the present amount of O2 is responsible for the generation of OTA+ on the time scale of the above experiment. Therefore, storage of the aerated complex solutions with added Sc(OTf)3 would eventually lead to the complete oxidation of OTA, albeit in a matter of days. Formation of OTA+ in the triad proceeded 2.6 times faster than in the dyad. Thus, all continuous irradiation experiments were conducted in aerated, as well as de-oxygenated complex solutions.

III.6.2 Irradiation of Aerated Samples in Presence of Sc3+ Figure III.49 depicts the UV-Vis absorption difference spectra after irradiation of Ru– OTA (Figure III.49 a) and AQ–Ru–OTA (Figure III.49 b) with a continuous light source at 450 nm. The spectra are recorded on the aerated solutions. The negative absorbance between 1350 and 1700 nm is due to the presence of OTA+ already directly after sample preparation. 0.6

b)

a)

0.5

DOD

0.4 0.3 0.2 0.1 0 -0.1 -0.2 400

600

800

1000

1200

wavelength / nm

1400

1600

400

600

800

1000

1200

1400

1600

wavelength / nm

Figure III.49 Changes in absorption upon continuous irradiation of aerated solutions of a) Ru–OTA (blue trace: 10 min; cyan trace: 3 h) and b) AQ–Ru–OTA (cyan trace: 30 min) in MeCN. λexc = 450 nm.

Spectra were recorded after 1, 2, 3, 4, 5, 10, 15, 30, 60, 120, and 180 min for Ru–OTA. Formation of OTA+ is complete after a little less than 10 min. After exactly 10 min, slight spectral indication of OTA2+ is already observable from the small shoulder at 1150 nm in the blue trace. After a total irradiation time of 3 h (cyan trace in Figure III.49 a), no more spectral changes are observed. However, compared to the extinction coefficients from the chemical oxidation experiment (Table III.6) the dyad system has rather reached a steady state, not complete formation of OTA2+, in the irradiation experiment. After 3 h, 70% of OTA+ is formed. The missing 30% might also be explained by the amount of OTA+ present prior to the irradiation experiments (see

96

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes “Sample Preparation”). Twofold oxidation of OTA was not expected for Ru–OTA, due to lack of an acceptor able to accommodate two electrons. [Ru(bpy)3]2+ is unlikely to accept two electrons. This again leads to the conclusion that O2 is the terminal electron acceptor also in the irradiation experiments. Nonetheless, irradiation of the Ru–OTA solution does accelerate the electron transfer reaction from OTA to O2 significantly. Within 10 min the 3.3-fold amount of OTA+ is formed in the continuous irradiation experiment as is in the same time in the reference experiment from above. Table III.6 Extinction coefficients in M–1cm–1 for OTA+ and OTA2+ species. Percentages are given in parentheses.

OTA+ OTA2+

chemical oxidation

Ru–OTA

AQ–Ru–OTA

17300 (100%) 59500 (100%)

12100 (69%) 56100 (94%)

– 51200 (86%)

In case of AQ–Ru–OTA, the spectra were recorded after irradiation times of 1, 2, 3, 4, 6, 8, 10, 15, and 30 min in aerated MeCN (Figure III.49 b). Already after 1 min of irradiation, formation of OTA+ is finished and a band at 1135 nm indicates the subsequent formation of OTA2+. After a total irradiation time of 30 min the band at 1130 nm for twofold oxidized OTA is present and further irradiation does not result in any more spectral changes. Comparing the extinction coefficient of OTA2+ after 30 min to the one obtained in the chemical oxidation experiment (Table III.6), indicates formation of 86% twofold oxidized OTA. This again points towards the system in a steady state. In this respect, the photoinduced formation of OTA2+ in the triad proceeds six times faster than in the dyad. Due to the presence of O2 in both experiments, this acceleration of the OTA oxidation is most likely due to the AQ acceptor in the triad. Formation of OTA2+ in the triad could in principle also yield twofold reduced AQ after two successive electron transfer steps. In order to facilitate the identification of bands due to AQ2–, the OTA2+ spectrum obtained through chemical oxidation was subtracted from the final spectrum from Figure III.49 after normalization of the bands at 1130 nm. This was done with the dyad, as well as the triad data and the resulting difference spectra are depicted in Figure III.50. For Sc3+-stabilized twofold reduced AQ a similar band shift as for its strongly hydrogen-bonded or even protonated counterpart is expected. Therefore, a band at 400 nm is expected for twofold reduced AQ in the difference spectrum of the triad. The traces depicted in the bottom half of Figure III.50 for Ru–OTA (blue) and AQ–Ru–OTA (cyan) do exhibit bands at 400 nm, however, this particular band is

97

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.6

DOD

0.4 0.2 0

DDOD

-0.2 0.2 0.1 0 -0.1 -0.2 400

600

800

1000

1200

1400

1600

wavelength / nm

Figure III.50 Top: Normalized spectra of Ru–OTA (blue) and AQ–Ru–OTA (cyan) in aerated MeCN solutions. Chemically generated OTA2+ spectrum in black. Bottom: Difference spectra of the dyad (blue) and triad (cyan) after subtraction of the OTA2+ spectral signatures.

present in both spectra. Furthermore, the spectra look very similar in general, and no band can be assigned exclusively to AQ in any reduction state. The lack of spectral evidence for reduced AQ is attributable to the present O2. The solubility of O2 in MeCN is ≈ 8.1 mM144 and its one-electron reduction potential is at −1.25 V vs. Fc+/0 in pure MeCN.145 Once formed superoxide (O2 –) anion was previously found to be thermodynamically very stable in MeCN, especially in presence of metal ions such as Sc3+.135,138 The mechanism leading to the formation of OTA2+ as evident from the spectra above, however without the appearance of reduced AQ related bands, may be interpreted as an AQ-mediated reduction of O2 (Scheme III.16). Sc3+ AQ-Ru-OTA



Sc3+

AQ--Ru-OTA+

AQ-Ru-OTA+ O2

O2



AQ--Ru-OTA2+

AQ-Ru-OTA2+ O2

3+

Sc

O2 Sc3+

Scheme III.16 Proposed mechanism for stepwise oxidation of OTA in AQ–Ru–OTA in presence of O2.

The potentials for the first reduction of AQ (−1.26 V vs. Fc+/0) and the reduction of O2 (−1.25 V vs. Fc+/0) are basically the same. If the potential for the AQ reduction is much more positive than −1.25 V, electron transfer from AQ – to O2 would not occur. On the other hand, a much more negative reduction potential of AQ compared to O2 would most likely favor direct electron transfer from the reduced photosensitizer to O2. In the present case, however, the reduced photosensitizer (−1.66 V vs. Fc+/0) is able to reduce either AQ or O2. Short-term generation of AQ – as an electron transfer intermediate is likely, since intramolecular electron transfer from OTA to AQ is very fast and exhibits a driving force of −0.92 eV. Time-resolved studies in pure

98

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes MeCN in absence of Sc3+ gave a lifetime of 2.4 µs for AQ––Ru–OTA+. During this rather long lifetime, O2 is able do diffuse towards the reduced AQ acceptor which can then reduce it. In an electrochemical study the bimolecular electron transfer rate from 2-hydroxymethylanthraquione (2-HM-AQ) to O2 in MeCN solution was found to be 7 · 106 M–1s–1.145 2-HM-AQ exhibits a potential of −1.33 V for the first reduction which leads to a slightly higher driving force for electron transfer to O2 than with AQ. Nevertheless, a similar rate constant can be assumed for the electron transfer between the reduced AQ acceptor in the triad and O2, especially in presence of the scandium metal ions which strongly bind and therefore stabilize the superoxide anion.135,138 The presence of OTA2+ in the dyad with no attached two-electron acceptor can be explained by a similar mechanism as above (Scheme III.17). Ru-OTA

hν O2

Ru-OTA+ O2 Sc3+

Scheme III.17

hν O2

Ru-OTA2+ O2 Sc3+

Possible mechanism for the stepwise twofold oxidation of OTA in Ru– OTA upon irradiation.

Though, no AQ-mediated electron transfer can occur in Ru–OTA, the twofold oxidation of OTA is made possible by the acceptance of the electrons by the present O2. In presence of Sc(hmpa)33+, O2 was found to exhibit a reduction potential of only −0.2 V vs. Fc+/0 in propionitrile.138 The AQ-mediated electron transfer mechanism as discussed above is likely to be the reason for the sixfold slower formation of OTA2+ in the dyad. As a result, successful photoinduced hole accumulation on the OTA donor was performed both in AQ–Ru–OTA and Ru–OTA in aerated MeCN. Though, no charge is accumulated on the AQ acceptor, it serves as an electrocatalyst in the system, promoting charge separation. To determine the influence of the excitation wavelength on the kinetics of the OTA2+ formation, dyad solutions in aerated MeCN were irradiated at 430 and 500 nm, respectively. At 430 nm OTA+ shows a local absorption maximum that superimposes the ruthenium MLCT band, thus, irradiation at 430 nm also excites OTA+. This might favor undesired back electron transfer re-reducing OTA+. In contrast, no OTA+ absorption is observed at 500 nm and irradiation at this wavelength should not promote the back reaction. However, excitation of the ruthenium photosensitizer will only occur into the very tail of its 1MLCT band. As a result, this experiment might also give an insight in the photostability of the generated Ru–OTA+ species, the influence of OTA+ excitation on the nature of the photoproduct. The irradiation times were the

99

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes same as in Figure III.49.

0.6

b)

a)

0.5

DOD

0.4 0.3 0.2 0.1 0 -0.1 -0.2 400

600

800

1000

1200

wavelength / nm

1400

1600

400

600

800

1000

1200

1400

1600

wavelength / nm

Figure III.51 Continuous irradiation of aerated MeCN solutions of Ru–OTA in presence of an excess of Sc3+. Excitation occurred at a) 430 nm and b) 500 nm.

Directly observable from the spectra in Figure III.51 is the formation of OTA2+ in both cases albeit at different rates. Irradiation of Ru–OTA at 430 nm shows complete OTA+ formation after 10 min (blue trace in Figure III.51 a) and formation of twofold oxidized OTA is finished after 3 h (cyan trace in Figure III.51 a). The obtained spectra are identical to the ones obtained after irradiation at 450 nm and the timescales observed for the OTA2+ formation agree well. Irradiation at 500 nm on the other hand shows much slower photoinduced generation of OTA+ and the twofold oxidized species. This however, is attributed to the fact that the photosensitizer is barely excited at this wavelength.

III.6.3 Irradiation of Aerated Samples in Absence of Sc3+ As shown above present scandium ions lead to a significant positive shift of the O2 reduction potential from −1.25 V vs. Fc+/0 to −0.2 V. Therefore, oxygen reduction in absence of Sc3+ should be somewhat more difficult and strongly influence the electron transfer kinetics in the continuous irradiation experiments. Figure III.52 depicts the difference absorption spectra after continuous irradiation of aerated samples of Ru–OTA and AQ–Ru–OTA in pure MeCN at 450 nm. From the graphs in Figure III.52 it is directly observable that no OTA oxidation occurs in the triad or dyad solutions in absence of Sc3+. This leads to the conclusion that not only the presence of O2 is responsible for the formation of OTA+ and even OTA2+ upon continuous irradiation of the aerated MeCN solutions. Addition of scandium ions is crucial to enable electron transfer from OTA to O2, since they significantly reduce the reduction potential of oxygen.

100

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.6

b)

a)

0.5

DOD

0.4 0.3 0.2 0.1 0 -0.1 -0.2 400

600

800

1000

1200

wavelength / nm

1400

1600

400

600

800

1000

1200

1400

1600

wavelength / nm

Figure III.52 Continuous irradiation of aerated MeCN solutions of Ru–OTA (a) and AQ–Ru–OTA (b) at 450 nm in absence of Sc3+. Cyan traces: after 3 h.

III.6.4 Dark Reactions of Aerated Samples in Presence and Absence of Sc3+ From the experiments above it is known that electron transfer from OTA to O2 as the terminal electron acceptor only occurs in presence of scandium ions. In the following the influence of the light source was studied. It was determined whether the electron transfer reaction was only dependent on the presence of O2 and Sc3+ in the sample and also occurs as a dark reaction, or if it is indeed photoinduced. Figure III.53 depicts the difference spectra recorded on dyad (left) and triad (right) solutions in aerated MeCN in presence (top) and absence (bottom) of Sc3+. All reactions were kept in the dark during the entire duration of the measurements. The total reaction times in presence of Sc3+ are the same as for the irradiation experiments (III.6.2); 3 h in case of Ru–OTA, and 30 min for AQ–Ru–OTA. In presence of 0.02 M scandium ions (Figure III.53 a and b) OTA oxidation is observed in aerated solution even in the dark. After 3 h in the dark, the spectrum recorded on the dyad shows a shoulder at 1140 nm, indicative of OTA2+ formation. With respect to the extinction coefficient of 8300 M–1cm–1 at 1140 nm, 14% OTA2+ are formed in the aerated dyad solution after 3 h, even in absence of light. In the aerated AQ–Ru–OTA MeCN solution in presence of scandium ions, OTA2+ formation reached a steady state in the irradiation experiment after only 30 min (Figure III.49). In the respective dark experiment (Figure III.53 b) no twofold oxidized OTA was observed after the same reaction time. However, oxidation of 37% of the amine to OTA+ did occur even in absence of light. Continuing oxidation to OTA+ and further oxidation to OTA2+ in the triad on a longer timescale is likely. Therefore, electron transfer from OTA to O2 yielding the scandium ion-stabilized superoxide anion is not a photoinduced reaction. Nevertheless, irradiation of the sample solutions does accelerate OTA2+ formation. In accordance with the results from the continuous irradiation experiments of both,

101

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.6

b)

a)

0.5

DOD

0.4 0.3 0.2 0.1 0 -0.1 -0.2 400

600

800

1000

1200

1400

1600

400

600

wavelength / nm 0.6

800

1000

1200

1400

1600

1400

1600

wavelength / nm d)

c)

0.5

DOD

0.4 0.3 0.2 0.1 0 -0.1 -0.2 400

600

800

1000

1200

wavelength / nm

1400

1600

400

600

800

1000

1200

wavelength / nm

Figure III.53 Difference spectra recorded on aerated MeCN solutions kept in the dark. a) Ru–OTA + Sc3+ (blue trace = 10 min, cyan trace = 3 h), b) AQ–Ru–OTA + Sc3+ (cyan trace =30 min), c) Ru–OTA without Sc3+ (cyan trace = 3 h), and d) AQ–Ru–OTA without Sc3+ (cyan trace = 3 h).

aerated dyad and triad solutions (Figure III.52), in absence of Sc3+ no OTA oxidation is observed. Based on the results of the above discussed experiments, it is concluded that O2, not AQ, is the final electron acceptor in the continuous irradiation experiments. The electron is transferred from OTA to O2 leading to OTA2+ and superoxide anion formation in the samples. The latter is strongly stabilized by coordination of scandium ions. In absence of the scandium ions no amine oxidation is observed. The metal ions are essential for the electron transfer reaction since they significantly facilitate O2 reduction and stabilize the resulting radical anion. The electron transfer process is not photoinduced, but it can be accelerated by irradiation. Furthermore, an AQ-mediated electron transfer mechanism can be described for the continuous irradiation of AQ–Ru–OTA in aerated MeCN solution in presence of Sc3+. This is the reason for the significant acceleration of OTA2+ formation in the triad, as compared to the dyad.

III.6.5 Irradiation of De-Oxygenated Samples in Presence of Sc3+ Since the generation of twofold oxidized OTA in the aerated samples in presence of 2000 eq Sc3+ can be attributed to photo-accelerated electron transfer from OTA to O2, all measurements were repeated in de-oxygenated solutions. Formation of OTA2+ in

102

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes de-oxygenated solutions would then point towards accumulative charge separation in the purely molecular AQ–Ru–OTA triad. Figure III.54 depicts the difference spectra obtained after continuous irradiation of de-oxygenated MeCN solutions of Ru–OTA (a) and AQ–Ru–OTA (b) in presence of Sc3+ at 450 nm. Spectra were recorded after irradiation intervals of 1 h. 0.2

b)

a)

DOD

0.1

0

-0.1

-0.2 400

600

800

1000

1200

wavelength / nm

1400

1600

400

600

800

1000

1200

1400

1600

wavelength / nm

Figure III.54 Difference spectra after continuous irradiation of de-oxygenated MeCN solutions of Ru–OTA (a)) and AQ–Ru–OTA (b)) in presence of 0.02 M Sc3+ at 450 nm. a) blue trace = 9 h. b) blue trace = 2 h, cyan trace = 9 h.

Continuous irradiation of the dyad solution over 9 h shows continuous formation of OTA+. After 9 h (blue trace) no more spectral changes are observed. Therefore, in deoxygenated MeCN solutions no twofold oxidized OTA is formed, as is the expected behavior in the dyad, due to the lack of an acceptor for two electrons. However, roughly 4 · 10−9 mol (2.4 · 1015 molecules) residual O2 are present during the experiment. With 1.8 · 1016 Ru–OTA molecules in the sample and considering an error of one order of magnitude for the residual O2 content in the cuvette, both are present in a 1:1 ratio and formation of OTA+ in the dyad can be ascribed to the present oxygen. In AQ–Ru–OTA, CI over 2 h leads to the quantitative generation of OTA+ (blue trace in Figure III.54). Continued irradiation of the sample leads to further oxidation of OTA+ and formation of the OTA2+ related band at 1130 nm. Compared to the data from the chemical oxidation experiment 23% OTA2+ is formed in the above experiment. The amount of AQ–Ru–OTA is 3 · 10−8 mol and calculation gave 4 · 10−9 mol of O2. After single oxidation of all complex molecules, in 23% of them OTA is oxidized a second time. Roughly 4 · 10−8 mol oxidant are therefore necessary to produce the spectrum in Figure III.54 which is well in range of the calculated amount of residual O2. OTA2+ formation in absence of O2 requires the twofold reduction of AQ. The presence of AQ2– should be indicated by an absorption band around 400 nm in the difference spectrum recorded on AQ–Ru–OTA. In the dyad, on the other hand, no such band is expected. Figure III.55 displays the final difference spectrum after an irradiation period of 9 h of the de-oxygenated triad solution (black trace). The black spectrum

103

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

mDDOD

DOD

can be fitted considering a 23% contribution of the OTA2+ (cyan trace) and 47% of the OTA+ (blue trace) spectrum. In theory, 77% OTA+ are still present after 9 h, however, the 30% OTA+ already present prior to the irradiation experiments have to be considered. 0.15 0.10 0.05 0 -0.05 -0.10 -0.15 0.05 0 -0.05 -0.1 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.55 Top: Difference spectra after 9 h of irradiation (black), of 47% OTA+ (blue), and 23% OTA2+ (cyan). Bottom: Difference spectrum after subtraction of the oxidized OTA contributions.

Though, the band maxima correlate very well in the NIR region, below 800 nm this fitting technique leads to unreliable results. The calculations above show that formation of OTA+ as well as twofold oxidized OTA can be completely ascribed to the residual O2. Therefore, the lack of a spectral signature for twofold reduced AQ in the difference spectrum can be taken as evidence for O2 being the final electron acceptor, as proposed above.

III.6.6 Irradiation of De-Oxygenated Samples in Absence of Sc3+ In absence of Sc3+ no formation of OTA+ or twofold oxidized OTA is expected. Investigation of the aerated samples showed that OTA oxidation is dependent on the presence of scandium ions in the samples and does not occur without them. Figure III.56 depicts the difference spectra obtained after irradiation of de-oxygenated MeCN solutions of Ru–OTA (Figure III.56 a) and AQ–Ru–OTA (Figure III.56 b) in absence of scandium metal ions. Irradiation of the dyad solution over 9 h seemed to induce slight degradation of the sample. This finding is not too surprising, considering the irradiation period. Furthermore, bimolecular processes can occur on this timescale that are not that easily tracked. In the case of AQ–Ru–OTA, minimal oxidation of OTA (6%) is observed even in absence of Sc3+ which was not expected in the CI experiment. Though photoinduced charge separation in the triad is the goal of the CI studies, the charge-separated state lifetime in pure, de-oxygenated MeCN (2.4 µs) was deemed

104

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.2

b)

a)

DOD

0.1

0

-0.1

-0.2 400

600

800

1000

1200

wavelength / nm

1400

1600

400

600

800

1000

1200

1400

1600

wavelength / nm

Figure III.56 Difference spectra after continuous irradiation of de-oxygenated MeCN solutions of Ru–OTA (a) and AQ–Ru–OTA (b) in absence of Sc3+ at 450 nm.

too short to accumulate enough AQ––Ru–OTA+ to detect on a UV-Vis timescale. The emission of the light source in the continuous irradiation experiments amounts to 5 photons per molecule per minute. Irradiation of the sample over 1 h then gives a total of 300 photons per molecule.

III.6.7 Dark Reactions of De-Oxygenated Samples in Presence and Absence of Sc3+ In presence of 0.02 M Sc3+ AQ–Ru–OTA shows 21% OTA+ formation after 9 h in deoxygenated MeCN solution even in the dark. Thus, after 9 h in the dark 6.3 · 10−9 mol of the triad are oxidized. In the analog dark experiment in the Ru–OTA dyad, 24% (7.2 · 10−9 mol) of OTA+ is formed. Within the accuracy of the calculations, the amounts of OTA+ formed in the de-oxygenated dyad and triad solutions in presence of Sc3+ are the same. Furthermore, the amount of OTA+ can be explained by the residual O2 (3.8 · 10−9 mol) in the de-oxygenated sample. The amount of residual oxygen in the Schlenk cuvettes is strongly dependent on the velocity of sample preparation and tightness of the cuvette. The final series of spectra displayed in Figure III.57 clearly shows that OTA is not oxidized in Ru–OTA in absence of Sc3+. Compared to the data obtained from the Sc3+-free de-oxygenated dyad solution that was irradiated over 9 h, decomposition of the sample in the dark is not as severe. In the case of the triad solution extremely slight OTA+ (4%) formation is observed. The analog experiment performed under irradiation showed formation of 6% OTA+ after 9 h. Such a small quantity of singlyoxidized OTA even in the dark-experiment might be attributed to the 2.4 µs lifetime of AQ––Ru–OTA+ in pure, de-oxygenated MeCN and the fact that total light exclusion is not achievable. In contrast to the dark experiments on the aerated solutions, the

105

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 0.2

b)

a)

DOD

0.1

0

-0.1

-0.2 400

600

800

1000

1200

1400

1600

400

600

wavelength / nm 0.2

800

1000

1200

1400

1600

1400

1600

wavelength / nm d)

c)

DOD

0.1

0

-0.1

-0.2 400

600

800

1000

1200

1400

1600

wavelength / nm

400

600

800

1000

1200

wavelength / nm

Figure III.57 Difference spectra recorded on de-oxygenated MeCN solutions kept in the dark for the duration of 9 h. a) Ru–OTA + Sc3+, b) AQ–Ru–OTA + Sc3+, c) Ru–OTA without Sc3+, and d) AQ–Ru–OTA without Sc3+. The cyan traces correspond to the spectra recorded after 9 h in the dark.

sample compartment cannot be closed with its lid when the Schlenk-cuvette is used for the O2-free measurements. Cuvette and sample compartment are only covered by a black cloth that might some light to pass through.

III.6.8 Possibility of Photo-Degradation On such long timescales of irradiation, photo-degradation cannot be fully excluded. A study on triarylamine radical cations has shown that thermally relatively stable radicals may undergo photoinduced electrocyclization.146,147 Visible light irradiation of the radical cations induces formation of the corresponding carbazoles albeit with rather small quantum yields. The probability for setting the triarylamine radical cation back to its uncharged ground state is bigger than electrocyclization. Though the photochemical generation of carbazoles cannot be fully excluded because excitation at 450 nm occurs partly into an OTA+ absorption band, it is difficult to find conclusive spectral evidence for their presence. Neutral carbazoles usually exhibit ground state absorption below 400 nm, as does neutral OTA.148 Furthermore, spectral features for oxidized carbazoles are a band around 450 nm and a rather sharp band at 800 nm.149,150 Due to the small extinction coefficient of the bands (roughly 450 M–1cm–1) their detection in the present CI experiments is difficult. However, the very similar spectra of AQ–Ru–OTA in pure MeCN and after addition

106

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

extinction coefficient / 103 M-1cm-1

80 60 40 20 0 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.58 Absorption spectra of AQ–Ru–OTA in pure MeCN (black trace) and of the de-oxygenated, irradiated solution after addition of NEt3.

of the reductant NEt3 to the irradiated, de-oxygenated sample of AQ–Ru–OTA in MeCN indicate complete re-reduction of OTA in AQ–Ru–OTA (Figure III.58). Addition of NEt3 induces precipitation of a white solid and the sample color changes back from a greyish tone to yellow. The white precipitate is most likely Sc(OH) 3– 151 Though, 6 which is formed in presence of an excess of aqueous alkaline solutions. hydroxide formation is not possible in the used pure, dry and de-oxygenated MeCN, the added NEt3 had not been dried and freshly distilled prior to its use as the reductant. It is therefore possible that the used NEt3 has imported trace amounts of H2O which lead to the hydroxide formation with scandium. Nevertheless, also based on the above shown spectra, no clear statement can be made concerning the presence of carbazoles after irradiation. OTA oxidation in Ru–OTA and AQ–Ru–OTA is dependent on the presence of O2 in the sample. In de-oxygenated solutions, with only trace amounts of O2 present, formation of OTA+ is slower than in the aerated samples. Oxidation also only occurs in presence of scandium metal ions and can be further accelerated by visible light irradiation of the sample solutions. A a result, the underlying reason for OTA2+ formation in the CI experiments is the interplay between O2, light, and the scandium metal ions. No intramolecular charge accumulation takes place in AQ–Ru–OTA. The AQ acceptor merely acts as an electrocatalyst for the electron transport from the OTA to O2 as the final electron acceptor.

107

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

III.7 Towards Accumulative Electron Transfer in an AQ–Ru–OTA–Ru–AQ Pentad The following section will focus on electron transfer in an AQ–Ru–OTA–Ru–AQ pentad. By attaching an extra ruthenium photosensitizer we hoped to increase the chance of a second excitation of the same molecule. In addition to the two metal sensitizers, a second anthraquinone acceptor was attached. By the provision of two acceptors, the only charge accumulating unit is the OTA donor, therefore hopefully minimizing the electrostatic problems associated with the buildup of charge. Another major advantage of the presence of two AQ acceptors in the pentad is that neither Brønstednor Lewis-acids have to be added to enable twofold reduction of AQ by the reduced photosensitizer. As discussed in III.4 and III.2, addition of acids in either form seemed to lead to partial, premature oxidation of OTA.

III.7.1 Synthetic Pathway to AQ–Ru–OTA–Ru–AQ Preparation of the pentad was started with the exchange of one bromine atom on 1,3,5-tribromobenzene by TMS after treatment with 1eq nBuLi and TMSCl. Br

TMS O B O hex

a) Br

Br 89 %

Br

Br c)

Br-TMS-Br

37 %

hex O B O

hex b)

99 %

I

I

hex

B-hex-B

TMS

OB O

hex

hex

hex

BO hex O

B-hex-TMS-hex-B

Scheme III.18 Synthesis of the core molecule: a) nBuLi, TMSCl, Et2O, -78 °C, 3 h, RT, over night; b) bis-pinacolato diboron, KOAc, PdCl2(PPh3)2, DMF, 90 °C, 2 d; c) Na2CO3, Pd(PPh3)4, toluene/EtOH/H2O, reflux, 3.5 d.

During the preparation of AQ–bpy–OTA as ligand for the Ru-triad solubility problems were observed. Thus, for the more extended organic backbone of the pentad, bearing two bipyridines and AQ, each, a benzyl spacer bearing two hexyl substituents was chosen. By this measure the solubility of the final ligand should be facilitated. The spacer, B–hex–B, was prepared from 1,4-dihexyl-2,5-diiodobenzene via Suzukicoupling with bis-pinacolato diboron in DMF in almost quantitative yield. Br–TMS–Br and B–hex–B were coupled Pd-catalyzed in a toluene/EtOH/H2O mix-

108

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

TMS O 2

N O

Br

N

hex

hex

hex

BO hex O

+ OB O

AQ-xy-bpy-Br

B-hex-B

71 %

a) X hex

hex hex hex

N O

N

N

N

X = TMS

O

43 %

b) O

O X=I

56 %

c) OCH3

OCH3

N

N

H3CO

N

OCH3

hex

hex hex hex

N O

N

N

N

O

AQ-bpy-OTA-bpy-AQ O

O

d) 32 % OCH3

OCH3

N

N

H3CO

N

4+ 4PF6-

OCH3

hex

hex hex hex

N O

N

N

N

N

Ru N

N Ru

N

N

N

O

N N

O

O

AQ-Ru-OTA-Ru-AQ

Scheme III.19

Synthesis of the final ligand: a) Na2CO3, Pd(PPh3)4, toluene/EtOH/H2O, 100 °C, 31 h; b) ICl, DCM/MeCN, 0 °C/RT, over night; c) OTA, NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, reflux, 5 d; d) AgNO3, Ru–pre, CHCl3/EtOH, reflux, 4 d.

109

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes ture to give the core compound B–hex–TMS–hex–B as a colorless oil. Regarding this coupling step, a fourfold excess of B–hex–B and a rather long reaction time of 3.5 d provided the best yield of 37%. Furthermore, gravitational column chromatography was necessary in order to fully separate impurities containing boronic ester functionalities (Scheme III.18). The core molecule prepared in this manner was further coupled in a Suzuki-type reaction to 2 eq of the AQ–xy–bpy–Br sidearm prepared in III.4. This gave AQ–bpy–TMS–bpy–AQ as a yellow solid in 71% yield. Subsequent deprotection of the TMS group by reaction with ICl gave 43% of the desired AQ–bpy–I–bpy–AQ product after column chromatographical purification. Finally, Pd-catalyzed N-C coupling in toluene with the OTA donor gave the ligand AQ–bpy–OTA–bpy–AQ as a brown solid in 56% yield. (Scheme III.19) As a final step, AQ–bpy–OTA–bpy–AQ was reacted with 3 eq of Ru–pre in presence of 6 eq of AgNO3 to facilitate the complexation. All reactants were suspended in a de-oxygenated CHCl3/EtOH mixture in the dark and heated to reflux for 4 d. The precipitated AgCl was filtered from the cooled reaction mixture and washed with DCM. Repeated column chromatography on SiO2 with an eluent mixture of acetone/H2O/1% aqu. KNO3 and subsequent precipitation with sat. KPF6-solution gave the desired pentad as the PF6-salt in 32% yield.

III.7.2 Electrochemical Characterization

current / µA

In Figure III.59 the cyclic voltammogram of AQ–Ru–OTA–Ru–AQ (bottom) is depicted in comparison to Ru–ref (top). Both voltammograms were recorded in Arpurged MeCN solutions with different internal standards. In case of Ru–ref, Fc+/0 was added and Me10Fc was used for the pentad. The sweep rate was always 0.1 V/s and 0.1 M TBAF6 was used as supporting electrolyte. 40 20 0 -20 -40 -60 40 20 0 -20 -40

*

*

-2

0 -1 +/0 potential / V vs. Fc

1

Figure III.59 Cyclic voltammograms of Ru–ref (black trace) and AQ–Ru–OTA–Ru–AQ (blue trace) in Ar-purged MeCN solutions. The internal references Fc+/0 for Ru–ref and Me10Fc for the pentad are marked by asterisks.

The oxidation potential for Ru3+/2+ in Ru–ref is 0.90 V vs. Fc+/0 and the potentials for the three ligand based reductions are −1.72 V, −1.92 V, and −2.17 V vs Fc+/0. These

110

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes values agree well with the ones reported in literature.36 All potentials obtained for the pentad are similar to the ones observed in the AQ–Ru–OTA triad. The stepwise oxidation of OTA to OTA3+ occurs at −0.01 V, 0.23 V, and 0.65 V vs Fc+/0 and the waves are less pronounced than in AQ–Ru–OTA. A reversible wave for the ruthenium oxidation is observable at 0.91 V. All bpy based reductions at −1.67 V, −1.92 V, and −2.20 V vs Fc+/0 agree well with those of the free ruthenium reference complex. As for the ruthenium triad, no shift in the photosensitizer-based redox potentials is observed in the pentad despite the attachment of the very electron-donating OTA. In the present case the distance between photosensitizer and OTA is too large for the latter to influence the reduction of Ru(bpy)32+. The first anthraquinone reduction at −1.25 V vs Fc+/0 is at the same potential as for the triad and coincides well with previously reported values.90 When the anthraquinone is reduced a second time to AQ2– (≈ −2 V) the same electrode effect was observed for the pentad as for the triad. The very sharp signal might be attributed to unwanted adsorption and stripping on the electrode which, in turn results in the irreversible reduction potentials. Appearance of the wave at −1.43 V might be due to electrochemical side-reactions and degradation reactions. Due to the more or less identical redox potentials of AQ–Ru–OTA and the pentad, similar thermodynamics are expected. Indeed, the first electron transfer step resulting in the formation of CSS1 (AQ–Ru–OTA+–Ru–AQ–) gives a value of −0.88 eV cal0 = e (−0.01 V − (−1.25 V )) − E . The triplet energy culated with the formula ∆GET 00 2+ 0 of [Ru(bpy)3] (E00 ) is 2,12 eV. ∆GET for the first electron transfer step in the triad was −0.92 eV. Formation of the second charge separated state AQ––Ru–OTA2+–Ru–AQ– 0 = e ((−0.01 V + 0.23 V ) − (−1.25 V + (−1.25 V ))) − E ) is exergonic by 0.6 eV (∆GET 00 after absorption of a second photon.

III.7.3 Optical Spectroscopic Characterization UV-Vis spectroscopic investigation of the pentad in MeCN solution gives a very similar spectrum to AQ–Ru–OTA. However, with an overall larger molar extinction, attributable to an even more extended conjugated organic ligand system. The pentad exhibits the typical 1MLCT band around 450 nm and the very distinct, bipyridine π* transition band at 288 nm.36 The increased absorption beligand-centered π tween 300 and 400 nm originates from OTA and AQ absorptions, which is also observed in the AQ–Ru–OTA triad. Compared to the free photosensitizer, the emission maximum of the pentad is redshifted by 7 nm to 610 nm. The emission intensity is quenched by 93%, indicating nonradiative deactivation of the excited photosensitizer. Regarding the thermodynamic evaluation above, deactivation most likely occurs via electron transfer. Triplet energy

111

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes 200 150

0.8 0.6

100

0.4

50

0.2

relative intensity a.u.

extinction coefficient / 103 M-1cm-1

1.0

0

0 300

400

500

600

700

800

wavelength / nm

Figure III.60 Absorption (solid trace) and steady-state emission (dashed trace) spectra of AQ–Ru–OTA–Ru–AQ in blue. The spectra of Ru–ref are shown for comparison (black traces). Complexes were measured in de-oxygenated 10−5 M MeCN solutions. Emission spectra were corrected for differences in absorption at 450 nm and normalized to the emission of Ru–ref.

transfer between the excited metal center (ET = 2.12 eV) and OTA (ET = 3.32 eV) can be excluded, because this would be a highly exergonic transfer step. The same holds true for energy transfer quenching of the excited photosensitizer by AQ (ET = 2.72 eV).123 Conclusive evidence for electron transfer in the pentad will be given below.

III.7.4 Photophysical Behavior in Different Solvents In addition to measurements in dry, de-oxygenated MeCN solution, the photoinduced electron transfer behavior of the pentad was investigated in an MeCN/H2O (v/v 1:1) solvent mixture, as well as in pure HFIP (hexafluoroisopropanol). Though not intrinsically necessary for photoinduced hole accumulation on the OTA donor, hydrogen-bond donors might facilitate it. Already in the MeCN/H2O solvent mixture AQ– should be stabilized by hydrogen-bond donation from H2O. In HFIP, a strongly hydrogen bond-donating solvent, the formed hydrogen bonds after photoexcitation should be even stronger.88 Usage of hydrogen bond-donating solvents, in contrast to the addition of acids, was expected not to lead to premature oxidation or protonation of OTA. The absorption spectra depicted in Figure III.61 were recorded in MeCN (black), MeCN/H2O (blue), and HFIP (cyan). No significant differences are observable in the spectra recorded in MeCN and MeCN/H2O solution. The noise between 1400 and 1500 nm is due to the first-overtone O-H stretching band of H2O.152,153 Contrary to the expectation, dissolution of the pentad in HFIP already leads to changes in the absorption spectrum. A slight hypsochromic shift is observed for the 1MLCT band of the photosensitizer and π* band a more pronounced two-band character is observed. In addition, the π is also blue-shifted from 288 nm to 282 nm. Furthermore, the broad absorption band with a local maximum at 325 nm which is a superposition of absorption bands for

112

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes OTA and AQ shows two local maxima at 316 and 350 nm in HFIP. This splitting of the initial band in HFIP might indicate interactions of the solvent with OTA.

extinction coefficient / 103 M-1cm-1

250 200 150 100 50 0 400

600

800 1000 1200 1400 wavelength / nm

1600

Figure III.61 Absorption spectra recorded on de-oxygenated, 10−5 M solutions of AQ– Ru–OTA–Ru–AQ in various solvents. MeCN = black, MeCH/H2O = blue, and HFIP = cyan.

In addition to the band splits and shifts in the absorption spectrum of AQ–Ru–OTA– Ru–AQ in pure, de-oxygenated HFIP, spectral indications for OTA+ are present at 736 and 1310 nm. When compared to the spectra obtained by chemical oxidation to OTA+ with Cu(ClO4)2, approximately 77% of OTA seem to be oxidized prior to the transient experiments. An extensive study by Gupta et al. has shown the presence of hydrogen bonds between HFIP and quinones in their charge-neutral form.131 Hankache et al. have provided clear evidence for hydrogen bond-strengthening between HFIP and AQ –, once the radical anion was formed after photoexcitation of a Ru–AQ complex. Furthermore, bonding of HFIP to charge-neutral AQ resulted in an increased drivingforce for the intramolecular electron transfer from the ruthenium metal sensitizer to AQ.88 In this respect, there is most likely hydrogen-bonding between charge-neutral AQ–Ru–OTA–Ru–AQ and the solvent HFIP. As a result, the driving-force for the first charge-separating step might have been increased initiating thermal electron transfer, leading to the partial, premature oxidation of OTA.

III.7.5 Time-Resolved Experiments in Different Solvents Luminescence Kinetics

The results of the steady-state emission experiment are supported by the kinetic study, showing complete quenching of the photosensitizer luminescence. The emission lifetime of the pentad is slightly increasing from MeCN (≈ 8 ns, kQ ≈ 108 s–1) and MeCN/H2O (≈ 10 ns, kQ ≈ 108 s–1) to HFIP (14 ns, kQ = 7.02·107 s–1). However, these values are very close to the instrumental detection limit. In the following, the nature and the lifetimes of the formed photoproducts are investigated and discussed in detail.

113

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

relative intensity a.u.

1

0.1

0.01

0.001 0

20

40

60

80 100 time / ns

120

140

Figure III.62 Emission decays of AQ–Ru–OTA–Ru–AQ in de-oxygenated MeCN (black), MeCN/H2O (blue), and HFIP (cyan). Excitation occurred at 532 nm and detection at 610 nm. Identification of the Photoproducts

mDDOD

D

The transient absorption spectra of the pentad in de-oxygenated solutions of MeCN (black), MeCN/H2O (blue), and HFIP (cyan) are depicted in the bottom half of Figure III.63. The sharp signal at 1064 nm is due to second order scattering light from the laser beam. All spectra shown were recorded at very high laser powers of 75 mJ. High excitation intensities should promote formation of the second charge-separated state and were therefore chosen for comparison of the transient spectra. More detailed power-dependency studies will be discussed later. In the visible range from 300 to 800 nm the transient absorption spectra closely resemble those obtained for AQ–Re–OTA (III.2.5) in presence of different acids. The black spectrum recorded in pure MeCN shows local absorption maxima at 410 and 565 nm, as well as a broad band stretching over the NIR region with an approximate maximum at 1320 nm. The features at 410 and 1320 nm correspond to the spectrum of OTA+ which is shown as a blue trace in the top half of Figure III.63. The additional band at 565 nm is attributed to the singly reduced AQ. Therefore, the main photoproduct in pure MeCN is AQ–Ru–OTA+–Ru–AQ–. 0.6 0.4 0.2 0 -0.2 40 20 0 -20 -40 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.63 Top: Spectral characteristics of OTA+ (blue) and OTA2+ (cyan) after chemical oxidation in MeCN. Bottom: Transient absorption spectra of AQ–Ru– OTA–Ru–AQ in MeCN (black), MeCN/H2O (blue), and HFIP (cyan). All spectra were measured of 10−5 M de-oxygenated solutions.

114

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes In the MeCN/H2O mixture a shift is observed in the visible region for the formed AQ –. It will most likely be hydrogen-bonded by the present H2O molecules, resulting in a band-shift from 565 to 500 nm.92 Furthermore, formation of the second charge-separated state in the transient experiment would lead to each AQ carrying one electron. In this respect, the intensity of the band corresponding to the reduced AQ is expected to be twice as intense as in the time-resolved experiments on AQ–Ru– OTA. This cannot be derived directly from the above shown spectrum and has to be investigated with power-dependent measurements, discussed later. The band around 400 nm is slightly red-shifted to 415 nm compared to pure MeCN, yet still indicative of OTA+. In the NIR region a band with a maximum at 1260 nm is observed. When considered individually and compared to the spectrum in pure MeCN and to the spectral signature of OTA2+ it seems to indicate the presence of OTA2+ in the photoproduct. However, the maximum is off by 130 nm, when compared to the spectrum obtained after chemical oxidation of OTA by Cu(ClO4)2 which shows a maximum at 1130 nm. Though, the general band shapes are similar, the maxima do not fit and the band at 415 nm indicates the presence of the singly oxidized OTA. The dropoff at 1400 nm, leading to the assumption of an OTA2+ band can be better explained by the presence of H2O. The absorption spectrum of the pentad in MeCN/H2O (blue trace) in Figure III.61 shows noise between 1400 and 1500 nm due to an increased absorption of H2O in this region. As a result, the NIR band in the transient spectrum is cut off at 1400 nm giving the illusion of a change in the band shape, although it is actually consistent with OTA+. The most likely photoproduct in the MeCN/H2O solvent mixture is the singly charge separated AQ–Ru–OTA+–Ru–AQ– with hydrogen bonds between AQ – and H2O. The overall intensity of the transient spectrum obtained from the HFIP solution is lower than in the other solvents. This might be attributable to the fact that oxidized OTA was already present before the photoexcitation experiments in a non-negligible amount. This behavior was also observed for AQ–Ru–OTA in presence of an excess of certain strong Lewis acids. The obtained spectrum in the visible region closely resembles the one measured of AQ–Re–OTA in presence of chloroacetic acid, which was assigned to AQH–Re–OTA+ (Figure III.23). Only one band is observed at 410 nm, and from 490 nm to longer wavelengths no separate bands are recognizable, but the optical density shows a steady rise. Doubly reduced and protonated AQ (AQH2) is known to show absorption around 400 nm with no further bands at longer wavelengths, which would explain the lack of the band around 500 nm.90,93 However, twofold reduced AQ is not expected in the pentad due to the presence of two AQ acceptors. Therefore, the more probable interpretation is the shift of the AQ – related band further into the

115

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes blue, superimposing with the OTA+ band at 400 nm. Protonation of AQ– (pKa = 5.3 in H2O)90 by HFIP (pKa = 9.3 in H2O)131 to form AQH is rather unlikely considering the thermodynamics. However, in the rather acidic HFIP, AQ – is very strongly hydrogen-bonded, possibly bordering on protonation. Although the signal in the NIR region is quite weak, a band with a maximum at 1000 nm can be observed. In this case, the dropoff at 1300 nm is not due to solvent absorption since HFIP does not show any absorption in this region. Yet again, the absorption maximum does not correlate with the OTA2+ band maximum at 1130 nm obtained in the chemical oxidation measurement. Furthermore, no increased absorption around 600 nm is observed for OTA2+. Consequently, and taking into account the knowledge gained in III.2.5, the photoproduct present in the HFIP solution is most likely AQ–Ru–OTA+–Ru–AQH. Kinetics of the Photoproducts in Different Solvents

Studies on AQ–Re–OTA have shown a remarkable increase in the lifetime of the generated photoproduct AQ––Re–OTA+ depending on the protonation degree. τ was increased from 205 ns for AQ––Re–OTA+ to 3.7 µs for AQH–Re–OTA+. Former studies on an AQ–Ru–TAA triad have also shown an increased lifetime of the chargeseparated state in presence of HFIP. In pure, de-oxygenated MeCN it was found to be 1.3 µs, which was prolonged to 3.0 µs in pure HFIP. In an equivalent AQ–Os–TAA triad even a 24.5-fold increase in lifetime was observed.154 In Figure III.64 the decays of the photoproducts of AQ–Ru–OTA–Ru–AQ in the different solvents as discussed above are depicted. Fitting of the decays gives the lifetimes of the singly chargeseparated states. relative intensity a.u.

1

1

a)

0.1

b)

415 nm 500 nm 780 nm

1

c)

410 nm 650 nm

0.1

410 nm 565 nm 790 nm

0.01 0

200

0.1

0.01 400

600

time / ns

800

1000

0

200

400

600

time / ns

800

1000

0

1

2

3

4

5

time / µs

Figure III.64 Transient absorption decays of the photoproducts in MeCN (a), MeCN/H2O (b), and HFIP (c).

AQ–Ru–OTA+–Ru–AQ– in pure, de-oxygenated MeCN has a lifetime of 160 ns. When changing from MeCN to the more hydrogen-donating MeCN/H2O mixture, the lifetime is increasing to 220 ns. This increase, though minor, may be attributed to the stabilization of the charge-separated state by hydrogen bonding from the wa-

116

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes ter molecules to the reduced AQ. In pure HFIP, a very strong hydrogen-bond donor, the lifetime is further increased. Here, the photoproduct shows a bi-exponential decay with lifetimes of 273 ns and 3.1 µs. The longer lifetime is very similar to the one obtained for AQ––Ru–TAA+ in pure HFIP and might therefore be the real lifetime of the photoproduct in the present experiment.154 Appearance of two species in the transient absorption experiments, indicated by bi-exponential decays, has been observed before in cases, where OTA was prematurely oxidized (III.4.7 and III.4.7). Table III.7 Gutmann acceptor numbers (AN)155 and dielectric constants ()156 of the solvents used. H2O represents the MeCN/H2O mixture.

solvent

AN

MeCN H2O HFIP

18.9 54.8 66.3



τ / ns

0.460 160 1.000 220 1.068 273

The observed trend in the lifetimes correlates well with the expectations due to the polarity and hydrogen bond-donating strengths of the used solvents. As a measure for the hydrogen bond-donating strength, the Gutmann acceptor number (AN) can be applied.155 It was presented primarily as a measure for the Lewis acidity of a solvent. However, in presence of a Lewis base, such as the reduced AQ, the acceptor number can be interpreted as a measure for the hydrogen bond donor strength of the solvent. The acceptor numbers and dielectric constants of the solvents, as well as the lifetimes of the respective photoproducts are summarized in Table III.7. In case of a 1:1 mixture of MeCN and H2O, the actual hydrogen bond-donating solvent is H2O, which is why it is listed in the table. The AN values clearly show an increase in hydrogen bond donor strength when going from pure MeCN to a MeCN/H2O mixture and to HFIP. HFIP is a very strong hydrogen bond donor due to its fluoro substituents. Excitation Density-Dependent Transient Absorption Spectroscopy

In the AQ–Ru–OTA–Ru–AQ pentad, formation of the desired second chargeseparated state with two holes on the OTA donor and one electron at either one of the AQ acceptors requires two-photon excitation.110,111 Therefore, excitation powerdependent measurements were performed. For all solvents, transient absorption spectra were recorded at high (75 mJ), medium (45 mJ), and low (10 mJ) power outputs. For MeCN and HFIP, the difference spectra were constructed from the spectra recorded at high and low power. In the case of MeCN/H2O, the sharp signal of the laser beam was visible at 45 mJ and it was even more prominent in the measurement with 10 mJ.

117

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes Since the rather intense laser signal coincides with the AQ band around 500 nm, the medium energy spectrum was used to construct the difference spectrum. Prior to the subtraction of the low from the high energy spectra, they had to be normalized. For the power-dependent studies on AQ–Re–OTA this was done by multiplying the low energy spectrum with the quotient of high and low power. This however, was not applicable for the measurements on the pentad, presumably due to saturation effects at 75 mJ. Thus, the spectral intensities were matched at 478 and 1285 nm, which represent the isosbestic points between OTA+ and OTA2+. Singly and doubly oxidized OTA shows absorption in almost all of the spectral range, superimposing also with the spectral signature of reduced AQ. Therefore, using the isosbestic point was deemed the most logical choice to normalize the spectra. Normalization on the AQ signature would seem logical, upon successful two-photon absorption, however, the band was expected to double in intensity. 40

a)

mDOD

10 0 -10 -20 4 2 0 -2 -4

b)

20 0 -20 -40

mDDOD

mDDOD

mDOD

20

5 0 -5 -10

600

800 1000 1200 wavelength / nm

mDDOD

mDOD

400

20 15 10 5 0 -5 -10 4 2 0 -2 -4

1600

400

600

800 1000 1200 wavelength / nm

1400

1600

c)

400

Figure III.65

1400

600

800 1000 1200 wavelength / nm

1400

1600

a) MeCN, b) MeCN/H2O, and c) HFIP. Top: Intensity-dependent transient absorption spectra (cyan = high energy; blue = low energy). Bottom: Corresponding difference spectra.

The intensity-dependent measurements on the MeCN solution (Figure III.65 a)) did not reveal any spectral indication for OTA2+ formation. No increased absorption is observed around 600 nm, or in the NIR region at 1130 nm. The same can be observed for the measurement in MeCN/H2O. In HFIP, a band at 1000,nm can be observed. However, this maximum deviates too much from the expected absorption at 1130 nm to be explainable by the poor sensitivity of the NIR detector. In the above experiments performed with our nanosecond pump-probe setup, no

118

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes unmistakable indications for the twofold charge-separated state AQ––Ru–OTA2+–Ru– AQ– could be found. The shift of the NIR band to shorter wavelengths in MeCN/H2O and HFIP may correlate with the formation of OTA2+. However, no increased absorption at 600 nm is observed even at very high excitation powers. Furthermore, 77% of the OTA donor have already been oxidized in HFIP prior to the laser experiments. Therefore, even if OTA2+ was observed unmistakably, the underlying mechanism would have been a one-photon process.

III.7.6 Photophysical Behavior in Presence of Thiourea Analogous experiments to the time-resolved study of AQ–Ru–OTA in presence of an excess of N,N-diphenylthiourea were performed on the pentad. In the present case a twofold increased excess was used to see if the increased concentration of diphenylthiourea changes the nature of the generated photoproduct. In Figure III.66 the ground state absorption spectra of the pure pentad (black) and with addition of 2000 eq of diphenylthiourea (grey) are depicted. The large excess of diphenylthiourea leads to the artefact at 340 nm and the noise at smaller wavelengths. All samples were prepared as 10−5 M de-oxygenated MeCN solutions. The added 2000 eq represent the upper solubility limit of diphenylthiourea in MeCN.

extinction coefficient / 103 M-1cm-1

140 120 100 80 60 40 20 0 400

600

800 1000 1200 1400 wavelength / nm

1600

Figure III.66 Ground state absorption spectra of 10−5 M MeCN solutions of the pure pentad (black trace) and with 2000 eq diphenylthiourea (grey trace).

It is observable from the graphs that even the addition of a large excess of thiourea does not lead to spectral changes attributable to oxidized OTA. Therefore, eventual spectral signatures for twofold oxidized OTA in the transient experiments can then be credited to a two-photon process. After the ground state absorption measurements, the same sample solution was investigated with the nanosecond pump-probe setup. The initial photosensitizer luminescence is quenched completely to τ = 12 ns, within the accuracy of the laser setup. In the upper part of Figure III.67 the traces representing the photoproduct after excitation with a 532 nm laser pulse are depicted. Transient spectra of the de-oxygenated

119

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes

mDDOD

mDOD

pentad solution were recorded at two different excitation energies. The blue trace corresponds to the photoproduct formed after excitation with a low energy pulse of 10 mJ, which translates to 4 photons per molecule per pulse. For the cyan trace, the pulse energy was 60 mJ which represents a photon count of 26 per molecule per pulse. As a result, generation of the twofold charge-separated state AQ––Ru–OTA2+– Ru–AQ– should theoretically be possible, considering the thermodynamics, as well as the photon flux. 20 15 10 5 0 -5 4 2 0 -2 -4 400

600

800 1000 1200 wavelength / nm

1400

1600

Figure III.67 Top: Transient absorption spectra of AQ–Ru–OTA–Ru–AQ in presence of an excess of diphenylthiourea at low (10 mJ; blue trace) and high energies (60 mJ; cyan trace). Bottom: Difference absorption spectrum by subtraction of the blue from the cyan trace. The spectra were normalized to the isosbestic points (478 nm and 1285 nm) prior to subtraction. All measurements were performed in de-oxygenated MeCN solutions.

In the upper part of Figure III.67 three main bands are observable over the whole spectral range investigated in the transient experiments. The sharp signal at 1060 nm is due to scattering of laser light. The absorption maxima of the two bands located in the visible range of the spectrum do not shift markedly upon the power increase. They are located at 413 and 535 nm. With the information gathered in the previous section III.7.5 and in the chemical oxidation experiment, the bands can be assigned to singly oxidized OTA at 413 nm and slightly hydrogen-bonded, singly reduced AQ at 535 nm. In pure MeCN solution the AQ– band can be found around 560 nm, whereas the band is blue-shifted to 500 nm upon hydrogen bonding in the MeCN/H2O mixture. In presence of diphenylthiourea, the band is in between the two extrema, maybe due to weaker hydrogen bonding between AQ– and the thiourea derivative. In the NIR region, a broad band is observed at low and high excitation energies. The local maximum, however, is slightly blue-shifted from 1300 nm at a low laser power, to 1200 nm at high power. From the chemical oxidation experiments the absorption maximum of OTA2+ in the NIR is known to be at 1130 nm once the band is completely developed. Therefore, the band maximum in the NIR does not fully agree with the expected absorption maximum of twofold oxidized OTA. This, however might in part be attributable to the low sensitivity of the NIR detector. As a result, assigning the

120

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes high power spectrum to AQ––Ru–OTA2+–Ru–AQ– would be doubtful, especially with the noise in the visible region. 0.6

DOD

0.4 0.2 0 -0.2 400

600

800 1000 1200 1400 wavelength / nm

1600

Figure III.68 Stepwise chemical oxidation of AQ–Ru–OTA–Ru–AQ by addition of incremental amounts of Cu(ClO4)2. Blue trace: OTA+; cyan trace: OTA2+. The arrows mark the isosbestic points used for normalization of the transient absorption spectra.

The supervention of a second, possibly twofold charge-separated state formed at high excitation power should be indicated by a double-exponential decay of the transient absorption signals. Figure III.69 depicts the decays of the absorption signals at 413, 540, and 740 nm in presence of diphenylthiourea. relative intensity a.u.

1

413 nm 540 nm 740 nm

0.1

0.01 0

200

400

600

800

1000

time / ns

Figure III.69 Kinetic absorption decays of the transient absorption bands at 413, 540, and 740 nm after excitation with a laser pulse of 60 mJ in presence of an excess of diphenylthiourea.

Exponential fitting gives a lifetime of 159 ns for the photoproduct in presence of an excess of thiourea. In pure, de-oxygenated MeCN, AQ–Ru–OTA+–Ru–AQ– has a lifetime of 160 ns. With regard to this it can be concluded that the added diphenylthiourea has no lifetime-prolonging effect on the photoproduct. Double-exponential fitting gives lifetimes of 100 ns and 292 ns. These lifetimes are rather similar as is already expected from the shape of the decays. However, it is debatable if a double-exponential fitting of the kinetic absorption data would indeed be logical. To further elucidate the possible presence of trace amounts of a twofold chargeseparated state, a difference spectrum of the blue and cyan spectra was constructed (orange spectrum; lower half of Figure III.67). For this, the high- and low-energetic

121

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes spectra were normalized at the isosbestic points at 478 nm and 1280 nm, as derived from the chemical oxidation experiment. Again, this approach seemed most reasonable due to oxidized OTA absorption in the whole spectral range, hampering the clear assignment of one neutral point, solely attributable to OTA+. An advantage of the difference spectrum is the subtraction of OTA+ related signatures from the obtained high-energy spectrum, leaving only bands assignable to twofold oxidized OTA and reduced AQ. The orange difference spectrum in Figure III.67 exhibits a rather low signal-to-noise ratio in the visible part of the spectrum. Clear assignment of actual bands is hampered, but slightly increased absorption is observable at 400, 500, and 620 nm. Compared to the difference spectra in the chemical oxidation experiment, the absorptions at 400 and 620 nm might indicate OTA2+ formation. In addition, the increased absorption at 500 nm might indicate the presence of hydrogen-bonded AQ–. The assumption of present OTA2+ is supported by the NIR band with a maximum at 1100 nm. Though, the absorption maximum is expected at 1130 nm, the deviation might be attributed to the lower sensitivity of the NIR detector. It is less sensitive than the iCCD camera, used for transient absorption measurements from 300 to 800 nm, and collection of data in the NIR region occurs at intervals of 10 nm with a band width of 10 nm. As a result, twofold charge-separated AQ––Ru–OTA2+–Ru–AQ– might indeed be present in the transient experiment, however, to a very small percentage. Nevertheless, this assumption requires further investigation with faster, maybe more accurate transient techniques, especially in the NIR region.

III.7.7 Summary and Conclusion on Electron Transfer in AQ–Ru–OTA–Ru–AQ In the preceding section, the synthesis, electro- and photochemical characterization, and photophysical behavior of an AQ–Ru–OTA–Ru–AQ pentad was presented. By the provision of two ruthenium photosensitizers twofold excitation of the same pentad molecule should be facilitated. As for the AQ–Ru–OTA triad, a major advantage of the use of ruthenium polypyridyl sensitizers over their rhenium analogues lies in the excitation wavelength. With the 532 nm pulse used to excite the ruthenium metal center, laser powers up to 80 mJ were achievable. In contrast to the corresponding triad, the focus in the pentad was solely on hole accumulation on the OTA donor. In principle, each of the AQ acceptors would then carry one electron. This approach should allow us to avoid the addition of strong Brønsted or Lewis acids to facilitate the second reduction of AQ. More importantly, the added acids caused premature oxidation/protonation of OTA in AQ–Ru–OTA,

122

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes sometimes to a non-negligible amount. This problem was eliminated by the use of the pentad. Theoretically and thermodynamically, hole accumulation on OTA should already be possible in pure solvent, such as MeCN. However, to aid the second charge separation by stabilization of the AQ– intermediate through hydrogen bonds, investigations in an MeCN/H2O solvent mixture and pure HFIP were performed. In all cases, formation of the photoproducts occurred within the pulse duration of the laser (10 ns) and the main product was found to be singly charge-separated AQ–Ru–OTA+–Ru– AQ–. In pure MeCN it exhibits a lifetime of 160 ns, whereas the lifetime increased from MeCN/H2O (220 ns) to HFIP (273 ns/3.1 µs). This increase in the lifetimes is due to the hydrogen bond-induced stabilization of AQ–. In HFIP it goes beyond hydrogen-bonding, and the reduced acceptor is protonated to form AQH. Though, the formed photoproduct in HFIP might be optimistically interpreted as the twofold charge-separated state AQH–Ru–OTA2+–Ru–AQH, UV-Vis studies showed premature oxidation of OTA. Therefore, formation of OTA2+ in the transient experiment would be only a one-photon process. In all cases, an increase of the excitation power did not lead to clear signals indicating the presence of the second charge-separated state. With an excess of N,N-diphenylthiourea, which would selectively coordinate to reduced AQ, no spectral changes indicative of OTA+ prior to the transient experiments were observed. The lifetime of the observed charge-separated state, most likely being AQ–Ru–OTA+–Ru–AQ– was 159 ns. This is an identical value to the one in pure MeCN and addition of diphenylthiourea did not lead to prolongation of the lifetime. Though, no indication for the twofold charge-separated state was found in the transient absorption kinetics, the difference spectrum does indicate formation of twofold oxidized OTA, especially in the NIR region. Clear assignment of the bands in the visible range of the transient spectrum is difficult. Nevertheless, the band at 1100 nm hints at the presence of a small amount of twofold oxidized OTA. Though thermodynamically favored, formation of the second charge-separated state depends on the positive interplay of various parameters, as discussed in section II.4. Even if AQ––Ru– OTA2+–Ru–AQ– is formed during the transient measurements, it will most likely be present to a very small percentage, which may complicate its detection. For the future, investigation of electron transfer reactions in the pentad with different laser setups might provide a more clear proof of hole accumulation on OTA. With a faster technology, e.g. on the femtosecond timescale, it might be possible to track photoproducts whose formation and decay is too fast for our nanosecond setup. Maybe even more promising would be measurements using a pump-pump-probe technique, allowing for the sequential excitation of the sample with two temporally delayed laser

123

III Towards Photoinduced Charge Accumulation in d6 Metal Complexes pulses possibly also involving two different excitation wavelengths. Using this technique, the transient spectra after the absorption of one and then two photons can be obtained more accurately than with the increase of the laser power. Furthermore, transient IR-spectroscopy might allow a deeper insight into the exact nature of the AQ acceptor in the charge-separated states. Lastly, a setup for a more precise recording of the NIR region would allow for a clearer assignment of the band to OTA2+.

124

IV. Solvent Dependent Emission of OTA–B(mes)2 Molecules Trirarylamine–triarylborane compounds have been studied extensively with regard to their emission properties.157–160 Organic push-pull systems are employed as dyes in solar cells, non-linear optics, anion sensing, and as emitters in organic light-emitting diodes.161,162 Due to this large area of possible applications, fundamental understanding and improvement of existing and creation of new, strong emitters is desirable. The oligo-triarylamine, at hand from the main research focus of this thesis, was further employed in push-pull systems. The main aim was to achieve superior charge transfer (CT) properties due to the remarkable electron donating abilities of OTA. In order to investigate the influence of various parameters on the emission behavior of such D–A compounds, a series of OTA–B(mes)2 molecules has been synthesized (Scheme IV.1). They vary in the donor and acceptor strength, the distance between D and A, and the position of A relative to D. OCH3

OCH3

H3CO

H3CO

N

H3CO

N B

N

N

B

N

B

B H3CO

H3CO

N

H3CO

TAA-short-B

N OTA-short-B2

OTA-short-B OCH3

OCH3

OCH3

OCH3

H3CO

H3CO

H3CO

N

N B

N

B

N

N

B

B H3CO

H3CO TAA-long-B

Scheme IV.1

H3CO

N

OCH3

OTA-long-B

N

OCH3

OTA-long-B2

Overview of synthesized donor-acceptor molecules.

125

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules

IV.1 Synthesis of Studied Molecules The compounds studied in this comparative study are donor–bridge–acceptor molecules comprised of an OTA donor and a boron dimesityl acceptor covalently linked by a conjugated bridging unit. In Scheme IV.2 the reaction pathways for the synthesis of the different acceptor building blocks and the donor–bridge building block are summarized. By coupling of the donor part with the respective acceptor, the final compounds as depicted in Scheme IV.1 were obtained.

Br

Br

a)

B

Br

80 % Br-ph-B

Br

Br Br

69 %

Br

B

a)

a)

B

Br

69 %

Br

B

Br2-ph-B Br-ph-B2

Br

b)

Br

Br

TMS

100 % Br-ph-acetylene-TMS c) 74 % OTA

d)

OTA

TMS

91 % OTA-ph-acetylene-H

Scheme IV.2

OTA-ph-acetylene-TMS

a) nBuLi, −78 °C, 2-3 h, dimesitylboron fluoride, RT, 1 h to over night; b) trimethylsilyl acetylene, PdCl2(PPh3)2, CuI, NEt3, 100 °C, 30 min; c) OTA, NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, reflux, 20 h; d) TBAF, THF, RT, 1.5 h.

The boron acceptor parts used to synthesize the new D–A compounds were prepared by attaching dimesitylboron units following a bromine-to-lithium exchange on the respective di- or tribromobenzenes. All preparations were performed in a similar manner. To a solution of 1,4-dibromobenzene in Et2O at −78 °C, nBuLi was added dropwise and the cold solution was stirred for 3 h. Afterwards, a solution of dimesitylboron fluoride in Et2O was added in a dropwise fashion, and the reaction mixture was stirred at RT over night. Following the work-up procedure, the product was obtained as a white solid with yields as high as 80%. In case of Br–ph–B2 stepwise attachment of the B(mes)2 units to 1,3,5-tribromobenzene proved to provide

126

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules a far better synthetic route than the one-pot reaction. After the two separate reaction steps the doubly B(mes)2-substituted product was isolated as a white solid in an overall yield of 48%. For the synthesis of the D–A molecules in which the donor and acceptor were separated by a longer bridge (OTA–long–B and OTA–long–B2), an acetylene spacer was introduced and coupled to the donor side. The spacer was prepared by Sonogashira coupling of trimethylsilyl acetylene to 1-bromo-4-iodobenzene. OTA was then coupled Pd-catalyzed to Br–ph–acetylene–TMS and the resulting product OTA–ph–acetylene–TMS was deprotected by stirring in a THF solution with TBAF. The donor part obtained in this manner was then coupled to the three different boron acceptors via Sonogashira coupling. Synthesis of the reference compounds TAA–long–B and TAA–short–B followed the route depicted in Scheme IV.3. H3CO

H3CO

Br-ph-acetylene-TMS

a)

+

N

TMS

64 %

b)

N

c)

H

99 %

43 %

Dianisylamine H3CO

H3CO

TAA-ph-acetylene-TMS

TAA-ph-acetylene-H H3CO

H3CO pBr-ph-B

d)

+

N

N

B

B

93 % Dianisylamine

Scheme IV.3

H3CO

TAA-short-B

H3CO

TAA-long-B

Synthetic route to the references: a) NaOtBu, Pd(dba)2, (PtBu3, toluene, reflux, 30 h; b) TBAF, THF, RT, 1 h; c) PdCl2(PPh3)2, CuI, NEt3, 100 °C, 30 h; d) NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, reflux, 2 h.

Of these two references, TAA–long–B was already known, but not yet investigated in terms of its emission behavior.159 Instead, it was cyclotrimerized to yield a trisdonor–tris-acceptor hexaarylbenzene which was then studied in detail. Preparation of the donor part for TAA–long–B occurred in analogy to OTA–ph–acetylene–TMS. Br–ph–acetylene–TMS was reacted with dianisylamine in a Pd-catalyzed Suzuki-type reaction to give the intermediate product TAA–ph–acetylene–TMS which was subsequently deprotected by stirring in THF with added TBAF. Final Sonogashira coupling with CuI as co-catalyst in dry and de-oxygenated NEt3 gave the desired TAA–long–B as a yellow solid in an over yield of 27%. For the preparation of TAA–short–B, Br–py–B was coupled to dianisylamine giving a bright yellow solid in

127

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules 93% yield. In Table IV.1 the colors of the individual compounds, as well as the overall yields are summarized. All final compounds were characterized by 1H NMR and 13C NMR, ESI-MS, and elemental analysis. Table IV.1 Synthesized D–A compounds with overall yields and colors.

compound OTA–long–B OTA–long–B2 OTA–short–B OTA–short–B2 TAA–long–B TAA–short–B

color

yield

orange orange bright yellow orange yellow bright yellow

23% 13% 62% 37% 22% 74%

IV.2 Solvent-Dependent Charge Transfer Behavior In the following section the six prepared donor–bridge–acceptor fluorophores are investigated with regard to their charge-transfer absorption and emission behavior in solvents of various polarities. The studied compounds are split into two groups (short vs. long) according to their different bridge lengths. In the molecules with the short bridging unit D and A are only separated by one phenyl spacer, whereas in the longer series the bridge consists of a phenyl–acetylene–phenyl unit. Within the two groups, the compounds vary in their donor and acceptor strengths in the order of TAA– bridge–B, OTA–bridge–B, and OTA–bridge–B2. This allows for a direct comparison of absorption, electrochemical, and emission behavior depending on the D–A distance and the strength of D and A.

IV.2.1 Electrochemical Characterization Figure IV.1 displays the cyclic voltammograms of the studied D–A compounds. All measurements were performed under ambient conditions in dry, Ar-purged THF solutions. The scan rate was 0.1 V/s in all cases and 0.1 M TBAPF6 was added as supporting electrolyte. The waves marked by asterisks are attributed to electrochemical side reactions as they appear only after one oxidative sweep beyond 0.5 V. All displayed voltammograms were measured on solutions without added Fc+/0 as the internal reference. In the data evaluation the cyclic voltammograms were shifted ac-

128

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules cordingly and are therefore referenced to Fc+/0. Direct addition of ferrocene lead to even more prominent side products than the ones marked by asterisks. *

-20 20 b) 0 -20

current / µA

current / µA

20 a) 0

*

40 c) *

0

20 d) 0 -20

*

10 e) 0 -10 20 f) *

0 -20

-40 -3

-2.5

-0.5 -1.5 -2 -1 potential / V vs. Fc+/0

0

0.5

-2.5

-2

-1 -1.5 -0.5 potential / V vs. Fc+/0

0

0.5

Figure IV.1 Cyclic voltammograms of the six fluorophores in dry, Ar-purged THF in presence of 0.1 M TBAPF6. Left: a) TAA-short–B, b) OTA–short–B, and c) OTA–short–B2. Right: d) TAA–long–B, e) OTA–long–B, and f) OTA–long– B2. Potential sweep rates are 0.1 V/s.

Oxidation of TAA in TAA-short–B (Figure IV.1 a) occurs at a potential of 0.33 V vs. Fc+/0 which is in line with previously reported values for this triarylamine donor.112,113 The first two OTA oxidations in OTA–short–B (Figure IV.1 b) occur at −0.02 V and 0.15 V. Such low oxidation potentials are typical for OTA and were often observed in purely organic, as well as in transition metal complexes containing the oligo-triarylamine donor unit.77,96 Usually, OTA exhibits also a third quasi-reversible oxidation wave around 0.6 V vs. Fc+/0 which is not shown in the cyclic voltammograms. In the present study on the emission behavior of the triarylamine–triarylboron compounds only the first oxidation of the amine which occurs at the central N atom is of interest.94 It is the electron-richest position in the amine, most readily releasing an electron, and charge transfer is also expected to occur via the shortest pathway between donor and acceptor. Therefore, OTA is a stronger one-electron donor than its TAA analog by ≈ 0.35 eV. Reduction of the boron mesityl acceptor in TAA–short–B, which occurs at −2.75 V, agrees well with values reported earlier.163 Substitution of TAA by the stronger electron donor OTA leads to a slight positive shift in the reduction potential of the boron mesityl unit to −2.69 V. In presence of a second acceptor unit in OTA–short–B2 reduction is again shifted positively to −2.52 V vs. Fc+/0. The increase in reduction potential in OTA–short–B2 is due to the presence of two electron withdrawing boron mesityl groups which facilitate reduction of the acceptor. Furthermore, the boron acceptors are located in meta-position to the OTA donor in OTA–short–B2, as compared to OTA–short–B. As a result, D and A are electronically more decoupled and transfer of electron density from OTA to the boron acceptor is expected to be weaker. From the redox potential values one can calculate the free energy for charge transfer

129

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules 0 between donor and acceptor, according to ∆GCT = e ( Eox − Ered ). The calculated values, together with the redox potentials extracted from the cyclic voltammograms are summarized in Table IV.2. It is directly observable that the driving force for charge transfer in the short D–A compounds decreases with increasing donor and unchanged acceptor strength, and then further with increasing acceptor strength.

Table IV.2 Redox potentials in THF given in V vs. Fc+/0.

complex TAA–short–B OTA–short–B OTA–short–B2 TAA–long–B OTA–long–B OTA–long–B2

amine+/0 amine2+/+ borane0/– /V /V /V 0.33 −0.02 −0.09 0.29 −0.06 −0.06

0.15 0.17 0.12 0.16

−2.75 −2.69 −2.52 −2.37 −2.36 −2.48

0 ∆GCT / eV

3.08 2.67 2.43 2.66 2.30 2.42

The electrochemical data obtained for the compounds with the longer D–A distance (Figure IV.1 right) show very similar amine oxidation behavior to their shorter analogs. TAA oxidation occurs at 0.29 V for TAA–long–B and the first two OTA oxidations for OTA–long–B and OTA–long–B2 are at −0.06 V and around 0.14 V vs. Fc+/0. Reduction of the triarylborane units in TAA–long–B and OTA–long–B occurs at a roughly 0.4 V less negative potential than in their analogs with the short phenyl bridge. However, attachment of a second boron mesityl acceptor does not lead to a further increase of the reduction potential of OTA–long–B2. As expected from the redox potentials, no clear ascending or descending trend is observed for the free energy of charge transfer for the longer D–A compounds. The charge transfer behavior of the studied D–A compounds seems to be less affected by the donor and acceptor strength when D and A are separated over a longer distance.

IV.2.2 Absorption and Emission Behavior in Hexane For the comparison of absorption and emission behavior of the six compounds, hexane was chosen as solvent. It is the most apolar of all the solvents used and all compounds were readily soluble. Figure IV.2 depicts the absorption (solid traces) and emission spectra (dashed traces) of TAA–short–B (black), OTA–short–B (blue), and OTA–short–B2 (cyan). The absorption spectra of purely organic charge transfer compounds show characteristic charge transfer (CT) bands that can give information about the push-pull system by their location and intensity.160 For the short D–A sys-

130

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules tems the CT bands are assigned to the lowest-energy absorption bands at 383 nm for TAA–short–B and at 410 nm for OTA–short–B. In OTA–short–B2 the CT band is visible only as a shoulder between 400 and 500 nm with an approximate local maximum at 450 nm. All higher-energy absorption bands can be assigned to π π* transitions of the amine donors or the triarylboron acceptor. The shift in the CT bands to 0 lower energies correlates well with the calculated ∆GCT values (Table IV.2) and puts in evidence the influence of increased donor and acceptor strengths on the charge transfer behavior. Therefore, the optical absorption data support the interpretation of the electrochemical data that OTA is indeed a better electron donor than TAA. The decrease in molar extinction from TAA–short–B to OTA–short–B and to OTA–short– B2 can be explained by the decrease in electronic coupling between D and A. This becomes particularly important in OTA–short–B2 where OTA and the boron mesityl acceptors are linked in meta-position to each other. Electronic coupling is known to be significantly weaker between meta-substituents linked via a phenylene bridge than between para-substituents.164

1.0 0.8

60

0.6

40

0.4 20

0.2

0

relative intensity a.u.

extinction coefficient / 103 M-1cm-1

80

0 300

400

500

600

700

800

wavelength / nm

Figure IV.2 Absorption (solid traces) and emission spectra (dashed traces) of TAA– short–B (black, λexc = 370 nm), OTA–short–B (blue, λexc = 330 nm), and OTA–short–B2 (cyan, λexc = 345 nm). All spectra were recorded on 10−5 M hexane solutions under ambient conditions and the emission spectra were normalized. λexc is the respective excitation wavelength.

TAA–short–B, OTA–short–B, and OTA–short–B2 show emission maxima at 454, 518, and 587 nm, respectively, as can be seen in Figure IV.2. These wavelengths correspond to energies of 2.73, 2.39, and 2.11 eV and correlate well with the calculated charge 0 transfer free energies ∆GCT in Table IV.2. In case of the three compounds with the longer D–A distance, both the CT absorption as well as the emission maxima do not show a steady red shift (Figure IV.3). However, the emission maxima of 443 (2.80 eV) for TAA–long–B, 508 (2.44 eV) for OTA–long–B, and 462 nm (2.68 eV) for OTA–long–B2 do follow the trend predicted by the calculated charge transfer energies in Table IV.2.

131

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules 1.0 0.8

60

0.6

40

0.4 20

0.2

0

relative intensity a.u.

extinction coefficient / 103 M-1cm-1

80

0 300

400

500

600

700

800

wavelength / nm

Figure IV.3 Absorption (solid traces) and emission spectra (dashed traces) of TAA– long–B (black, λexc = 390 nm), OTA–long–B (blue, λexc = 345 nm), and OTA–long–B2 (cyan, λexc = 345 nm). All spectra were recorded on 10−5 M hexane solutions under ambient conditions and the emission spectra were normalized. λexc is the respective excitation wavelength.

Similar to the short D–A compounds, the absorption spectra of their longer analogs show π π* transitions below 350 nm and the CT bands at lower energies. The exact values of the CT band maxima are 397 nm for TAA–long–B and 417 nm for OTA–long–B. In case of OTA–long–B2, the CT band is only visible as a shoulder at 385 nm. The intensities of the CT bands decrease with increasing donor and acceptor strengths, however, the effect is much less pronounced than in the D–A molecules with the phenyl bridge.

IV.2.3 Solvatochromism In order to study the dependence of the CT absorption and emission on the solvent polarity, spectra of all six D–A compounds were recorded on 10−5 M solutions in five different solvents. The solvent polarity ranges from hexane (as the most apolar), toluene, Et2O, and THF to DCM as the most polar (Table IV.3). In solvents of higher polarity (e.g. acetone, DMF, and DMSO) the compounds showed very low luminescence quantum yields and unambiguous differentiation between emission originating from the studied compounds and artifacts was difficult. Furthermore, solubility was limited in highly polar solvents such as MeCN and MeOH. Therefore, those solvents were excluded from the studies. Table IV.3 Solvents and their polarity in terms of Reichhardt parameters ETN.156

solvent ETN

hexane

toluene

Et2O

THF

DCM

acetone

DMF

MeCN

MeOH

0.009

0.099

0.117

0.207

0.309

0.355

0.404

0.460

0.762

Figure IV.4 depicts the absorption spectra of all compounds in the five selected solvents. The short D–A molecules are depicted in the top half (a-c) and their longer analogs in the bottom part (d-f).

132

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules

extinction coefficient / 103 M-1cm-1

40

60

a)

c)

50 60

30 40

40

30

20

20 20

10 10 0

0

0 300

400

500 600 700 wavelength / nm

300

800

400

500 600 700 wavelength / nm

300

800

80 e)

80

30

60

60

20

40

40

10

20

20

40

extinction coefficient / 103 M-1cm-1

80

b)

d)

0 400

500 600 700 wavelength / nm

800

500 600 700 wavelength / nm

800

400

500 600 700 wavelength / nm

800

f)

0

0 300

400

300

400

500 600 700 wavelength / nm

800

300

Figure IV.4 Absorption spectra of a) TAA–short–B, b) OTA–short–B, c) OTA–short–B2, d) TAA–long–B, e) OTA–long–B, and e) OTA–long–B2 recorded on 10−5 M solutions in hexane (black), toluene (blue), Et2O (cyan), THF (red), and DCM (orange).

From the spectra it is directly observable that the location of the CT absorption band is not influenced by the different solvents. Strong polarity-dependence or solvatochromism is usually only observed for fluorophores that exhibit a dipole moment themselves.34 In their electronic ground state the investigated compounds are certainly somehow polarized due to their push-pull character, however, the effect is not strong enough to induce spectral shifts in the different solvents. Once the D–A molecules are excited, charge transfer from the amine moieties to the boron acceptors is induced and the dipole moment is significantly increased in the charge-separated state. As a result, a much stronger solvatochromic effect is expected in the emission spectra (Figure IV.5). All studied compounds show a distinct redshift of their emission with increasing polarity, as is expected for such push-pull systems.34,157,158,165 The positive solvatochromic shift in TAA–short–B between hexane and DCM amounts to 3050 cm–1 whereas in OTA–short–B it is only 2720 cm–1. Overall, the emission maxima for OTA–short–B are shifted further into the red, as was expected due to the lower charge transfer energy. OTA–short–B2 only exhibits significant emission in hexane. The longer analogs also display positive solvatochromism, however, OTA–long–B and OTA–long–B2 do not show significant emission in THF and DCM. Therefore, the shift between hexane and Et2O is taken as a measure for the solvatochromic shift in

133

relative intensity a.u.

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules 1.0 a)

1.0 b)

1.0 c)

0.8

0.8

0.8

0.6

0.6

0.6

0.4

0.4

0.4

0.2

0.2

0.2

0

0

relative intensity a.u.

400

500 600 700 wavelength / nm

0 400

500 600 700 wavelength / nm

400

1.0 d)

1.0 e)

1.0 f)

0.8

0.8

0.8

0.6

0.6

0.6

0.4

0.4

0.4

0.2

0.2

0.2

0

0 400

500 600 700 wavelength / nm

500 600 700 wavelength / nm

0 400

500 600 700 wavelength / nm

400

500 600 700 wavelength / nm

Figure IV.5 Emission spectra of a) TAA–short–B, b) OTA–short–B, c) OTA–short–B2, d) TAA–long–B, e) OTA–long–B, and e) OTA–long–B2 recorded on 10−5 M solutions in hexane (black), toluene (blue), Et2O (cyan), THF (red), and DCM (orange).

those compounds. The magnitudes of the solvatochromic shifts present themselves as follows: 3000 cm–1 for TAA–long–B, 2760 cm–1 for OTA–long–B, and 3720 cm–1 for OTA–long–B2. As a result, the reference compounds TAA–short–B and TAA–long–B show the strongest solvatochromism, with exception of OTA–long–B2.

IV.2.4 Difference between Ground and Excited State Dipole Moments In 1955 Lippert and Mataga have communicated their theory for solvent–fluorophore interactions. Though it does not consider specific solvent effects (e.g. hydrogenbonding, acid-base chemistry, CT interactions) it is a valuable method to analyze the solvent-dependent absorption and emission behavior of fluorophores.34,166,167 Equation IV.1 allows for the estimation of the difference between ground- and excitedstate dipole moments (∆µeg ) on the basis of the experimentally determined Stokes shifts in various solvents. hc · (νabs − νem ) =

vac hc · (νabs

vac − νem )·

2 · ∆µeg a30

2   e−1 η2 − 1 · − 2e + 1 2η 2 + 1

(IV.1)

In Equation IV.1, (νabs − νem ) is the Stokes shift that can be extracted from the absorption and emission data from Figure IV.4 and Figure IV.5, h is Planck’s constant,

134

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules

hc(nabs-nem) / cm-1

and c is the speed of light. Furthermore, e is the dielectric constant, η the refractive index of the solvent, and a0 is the Onsager radius. In Lippert-Mataga theory, the Onsager radius is defined as the solvent cavity in which the fluorophore is embedded, and for easier determination of this radius the fluorophores are assumed to be spheres.168 The solvent polarity is described by the so-called orientation polarizability ∆ f = (e − 1)/(2e + 1) − (η 2 − 1)/(2η 2 + 1). In this term, (η 2 − 1)/(2η 2 + 1) describes the reorientation of the electrons of the solvent molecules after the change in dipole moment on the CT compound. Electrons adjust basically instantly to the changed environment, hence the term is also called the high-frequency polarizability. (e − 1)/(2e + 1) describes the reorganization of the solvent molecules, which occurs on a much slower timescale than electron reorganization and is therefore denoted as low-frequency polarizability.34 8000 a)

8000 b)

7000

7000

6000

6000

5000

5000

4000

0

0.05

0.10

Df

0.15

0.20

0.25

4000

0

0.05

0.10

Df

0.15

0.20

0.25

Figure IV.6 Lippert-Mataga plots of the Stokes shift vs. the orientation polarizability ∆ f . a) TAA–short–B, b) OTA–short–B.

Figure IV.6 displays the plots of hc · (νabs − νem ) vs. ∆ f for TAA–short–B (a) and OTA–short–B (b). Linear regression fits of the data yield values of 10500 ± 3000 cm–1 (TAA–short–B) and 10000 ± 2300 cm–1 (OTA–short–B) for the slopes, which are the same within experimental accuracy. Ideally, the difference in ground- and excitedstate dipole moments for the investigated compounds can be estimated from the slope which is equal to ∆µ2eg /a30 . Determination of the Onsager radius is always prone to errors and these have an even larger impact on the calculated value for ∆µeg due to its dependence on the third power of a0.169 Furthermore, already the linear regression fits do exhibit a non-negligible error due to the relatively poor linear correlation (R2 = 0.81 for TAA–short–B and R2 = 0.86 for OTA–short–B) between the Stokes shifts and solvent polarity as depicted in Figure IV.6. Another source of error in the LippertMataga equation (IV.1) is the assumption of spherical fluorophores. The structural formula of TAA–short–B and OTA–short–B already imply the deviation from this assumption. Consideration of an elliptical solvent cavity would certainly be better suited for the present compounds.

135

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules

hc(nabs-nem) / cm-1

8000 7000 6000 5000 4000 3000 2000

0

0.05

0.10

Df

0.15

0.20

0.25

Figure IV.7 Lippert-Mataga plot of the Stokes shift vs. the orientation polarizability ∆ f for TAA–long–B.

The same general observations as discussed above can be made for TAA–long–B (Figure IV.7). Linear fitting leads to a slope of 18000 ± 3300 cm–1 which is substantially higher than the ones obtained for TAA–short–B and OTA–short–B. However, the uncertainty of the linear fit is even higher (R2 = 0.91) and the same errors regarding the estimation of the Onsager radius have to be considered here. Therefore, also in the longer D–A compound, derivation of an accurate value for ∆µeg from the plots would be questionable. The rather strong deviation of the experimentally obtained values from the ideal, expected linear behavior indicates the presence of more specific interactions between the fluorophores and the solvents. OCH3

H3CO

H3CO

N

N

N

B

H3CO

H3CO

B

N

OCH3

smaller Dµeg smaller a0

larger Dµeg larger a0

similar solvatochromism Figure IV.8 Illustration of increased Onsager radii in TAA–short–B (left) and OTA– short–B (right).

Nevertheless, a semi-quantitative statement regarding the change in ∆µeg upon an increase in donor strength in a system with a fixed D–A distance might be appropriate. Compared to TAA–short–B, substitution of TAA by OTA may be associated with an increase of the Onsager radius by 25% (Figure IV.8). This estimation is based on the increased molecular size of OTA–short–B. Though this increase of the compound

136

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules radius is quite small, application of hc · (νabs − νem ) = ∆µ2eg /a30 with the experimentally determined Stokes shift and the 25% increased a0 can mask an increase in ∆µeg of up to 40%. OTA–short–B does very likely exhibit a significantly stronger change in its dipole moment after excitation and charge transfer than TAA–short–B, due to the presence of the stronger electron donor OTA. Alas, this effect does not lead to stronger solvatochromism, because the effect of the stronger donor is outbalanced by the increase in a0. As a result, to create systems with very strong solvatochromism, the molecular size should be kept as small as possible with increasing donor and acceptor strengths.

IV.2.5 Luminescence Lifetimes and Quantum Yields The quantum yields and the corresponding luminescence lifetimes of the CT states of the D–A compounds TAA–short–B, OTA–short–B, and TAA–long–B were measured in hexane, toluene, Et2O, THF, and DCM under ambient conditions. For OTA–long– B and OTA–long–B2, determination of φ and τ was only possible in hexane, toluene, and Et2O, whereas OTA–short–B2 shows significant luminescence in hexane. All measurements were performed on 10−5 M solutions. Table IV.4 displays the summarized luminescence data. One general observation that can be made for all investigated compounds is a concomitant decrease in the luminescence quantum yield with decreasing CT energies of the excited states. The so-called energy gap law describes the dependence of the efficiency of non-radiative deactivation processes on the energy gap between the ground state and the luminescent excited state. With decreasing energy gap, non-radiative processes become more and more efficient.84 This is in line with the observed decrease in quantum yields for the D–A compounds in different solvents. The lack of correlation between the quantum yields and luminescence lifetimes indicates once more the presence of specific solvent–fluorophore interactions that lead to significant, highly solvent-dependent variations in efficiency for non-radiative and radiative deactivation processes. Interactions between solvent donor molecules and the triarylborane acceptors might open a thermal relaxation pathway to the excited states that becomes even more apparent as the polarity of the solvent increases. Compared to TAA–short–B, OTA–short–B, TAA–long–B, and even OTA–long–B, the two compounds bearing two boron mesityl acceptors exhibit much poorer luminescence properties. In the present study, the acceptor strength was increased by attachment of two triarylboranes. Due to the design of the D–A compounds this requires the two acceptors to be attached in meta-position to the OTA donor. As already stated above, this leads to a significant decrease in electronic coupling between D and A

137

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules Table IV.4 Luminescence quantum yields (φ) and lifetimes (τ in ns) of the fluorophores in the different solvents. Measured in aerated solvents under ambient conditions.

hexane

toluene

Et2O

THF

DCM

Compound

φ

τ

φ

τ

φ

τ

φ

τ

φ

τ

TAA–short–B OTA–short–B OTA–short–B2 TAA–long–B OTA–long–B OTA–long–B2

0.59 0.49 0.06 0.81 0.63 0.21

4.1 5.8 11.8 1.8 3.6 3.8

0.72 0.50

6.0 7.8

0.75 0.40

7.4 7.0

0.63 0.12

7.6 2.7

0.66 0.04

8.6 1.1

0.85 0.52 0.13

2.6 4.1 9.3

0.88 0.11 0.08

3.6 1.8 13.2

0.49

2.8

0.38

2.4

and therefore, in a decreased efficiency of the radiative decay rate for OTA–short–B2 and OTA–long–B2. Furthermore, if thermal relaxation of the excited CT states due to solvent interactions with the boron acceptor is indeed possible, despite the bulkiness of the triarylboranes, this effect is certainly increased in presence of two acceptors in OTA–short–B2 and OTA–long–B2. Since the four stronger luminescent compounds TAA–short–B, OTA–short–B, TAA– long–B, and OTA–long–B exhibit similar oscillator strengths, their radiative decay constants should also be similar in the used solvents. However, the quantum yields do differ, which might be attributed to differences in non-radiative excited state relaxation processes. In the OTA bearing D–A molecules, more atoms are present due to the more extended molecular system of OTA. Therefore, the number of available vibrational degrees of freedom is higher in OTA–short–B and OTA–long–B and allows for more efficient deactivation of the excited states. Comparing the short to the long D–A compounds, the additional ethynyl unit only exerts a small influence on the luminescence quantum yields, compared to the stronger OTA donor and the meta-coupling of the two boron acceptors. This small effect is only observable in the most apolar solvents hexane, toluene, and Et2O. Already in the weakly polar solvents THF and DCM, the effect becomes negligible.

IV.3 Summary The aim of this comparative study was the synthesis and physicochemical investigation of six D–A compounds (Scheme IV.1) with regard to their charge transfer and solvatochromic behavior. The structural differences in the compounds allowed for the study of the influence of donor and acceptor strengths, position of D with respect to A, and the bridging length on their luminescence behavior. Consideration of all

138

IV Solvent Dependent Emission of OTA–B(mes)2 Molecules experimental data results in some key findings. On the basis of the cyclic voltammetry data for the donor and acceptor moieties, the trends in charge transfer energies could be predicted for all unknown compounds. Use of OTA as a stronger electron donor compared to TAA did lead to the desired red-shift of the CT emission and a stabilization of the CT state. However, due to the increased size of the OTA-containing compounds, the Onsager radius was also increased. Therefore, the overall solvatochromic effect is comparable to the one in the TAA-containing compounds that exhibit a smaller dipole moment change between the ground and excited states. Increase of the acceptor strength by attachment of a second boron mesityl acceptor influences the CT properties negatively. The two triarylboranes require attachment in meta-position to the amine donors which leads to a decreased electronic coupling between D and A. This in turn leads to a significant decrease in the luminescence quantum yield. Based on these key findings, design guidelines for future amine–bridge–borane fluorophores are proposed: First, when aiming to enhance the CT properties of such push-pull compounds the donor and acceptor strengths should be increased without a concomitant increase of the Onsager radius, because these effects tend to annul each other. Second, very electron-rich arylamine donors do present themselves as perfect candidates for use in such push-pull systems, however, their size should be kept small. Third, the acceptor strength of triarylboranes should be increased by substitution of the aryls (e.g. fluorination), rather than by attachment of several acceptors in metaposition. Fourth, a high radiative decay rate of the CT state is necessary and can be achieved by strong electronic coupling between D and A. Therefore, a suitable bridging unit is necessary and p-phenylene and p-phenylene ethylene linkers show similarly positive behavior in this respect.

Photographs of the emission behavior of the strongest emitter TAA–short–B (left) and the weakest emitter OTA–short–B2 (right) in the used solvents.

139

V. General Summary The main aim of this thesis was the achievement of photoinduced charge accumulation in purely molecular, nanoparticle-free d6 transition metal sensitized compounds. All investigated compounds were comprised of various combinations of an OTA donor, an AQ acceptor, and either a rhenium or ruthenium polypyridyl photosensitizer. Despite the efforts of constant improvement of the molecular designs and the measuring parameters, no unequivocal evidence for twofold charge separation could yet be found in any of the investigated systems with the available pump-probe nanosecond setup. The studied ruthenium dyad, triad, and pentad allowed for significantly higher excitation densities of the laser beam than their rhenium analogs and are therefore preferable for future studies. Ru–OTA in presence of TCNB and AQ–Ru–OTA, as well as AQ–Ru–OTA–Ru–AQ in presence of diphenylthiourea displayed spectral features at high laser powers that do hint towards the presence of the second charge-separated states. Unambiguous bands could not be fully resolved with the used pump-probe nanosecond setup, but also cannot be fully attributed to noise or instrumental artifacts. Therefore, pump-pump-probe experiments using two temporally delayed excitation pulses, maybe even of different wavelengths, would certainly represent a major advantage and will be pursued in the near future. When performed with sufficiently high temporal resolution, < 10 ns, even very short-lived photoproducts might be detectable. Employment of the strong electron donor OTA in a side project, investigating pushpull systems with boronmesityl acceptors, did not lead to the desired superior charge transfer and solvatochromic properties. However, important guidelines regarding the development of future triarylamine-triarylborane emitters could be worked out.

141

VI. Experimental Part

VI.1 Analytical Methods General Methods All commercially available chemicals were used without further purification. Dry DCM, Et2O and THF were obtained from a solvent purification system of Innovative Technology. Dry toluene was bought from Sigma-Aldrich (crown cap; over molecular sieve) and used as received. The additional solvents hexane, acetone, DMF, and DMSO used for polarity dependent measurements were also bought from SigmaAldrich (crown cap; over molecular sieve). Thin-layer chromatography was performed on silica gel plates (60 F254) from Merck. Silica gel (40–63 µm, Silicycle) or activated, basic aluminum oxide (Brockmann I, 50–200 µm) from Acros was used for column chromatography. All given temperatures in the synthesis procedures are the set oil bath temperatures.

NMR Spectroscopy 1

H NMR spectra were recorded on a 250 MHz or 400 MHz Bruker Avance III, 13C spectra on a 500 MHz Avance III from Bruker. All shifts (δ) given below are in ppm and referenced to residual solvent peaks. Coupling constants are listed in Hz and the multiplicity is denoted as follows: singlet (s), broad singlet (bs), doublet (d), triplet (t), quartet (q), multiplet (m).

High Resolution Mass Spectrometry The high-resolution mass spectra were measured on a Bruker maXis 4G QTOF ESI spectrometer.

Elemental Analysis All elemental analyses were performed on a Vario Micro Cube from Elementar.

143

VI Experimental Part

X-ray Crystallography The crystal was measured on a Bruker Kappa Apex2 diffractometer at 123 K using graphite-monochromated Cu Kα radiation.

FT–IR Spectroscopy FT–IR spectra of the complexes in solid form were measured on a Bruker Alpha Platinum-ATR.

UV–Vis Spectroscopy All UV–Vis spectra were recorded on a Cary 5000 UV–Vis–NIR spectrophotometer.

Cyclic Voltammetry A VersaStat3–200 potentiostat of Princeton Applied Research was used for electrochemical experiments. The measurements were performed in dry, argon-saturated MeCN with tetrabutylammonium hexafluorophosphate (TBAPF6) as supporting electrolyte and using a conventional three-electrode setup. A glassy carbon disc was used as the working electrode and two silver wires served as the quasi-reference and the counter electrode, respectively. All cyclovoltammograms are referenced versus the ferrocene/ferrocenium (Fc+/0) redox couple.

Steady-state Luminescence, Lifetimes, and Quantum Yields Steady-state luminescence spectra were obtained by a Fluorolog-322 from Horiba Jobin-Yvon. Fluorescence lifetime measurements were done on a Hamamatsu Quantaurus–Tau C11367-11. Absolute photoluminescence quantum yields were measured on a Quantaurus–QY C11347-11 from Hamamatsu.

Transient Absorption Transient absorption was measured with an LP920-KS spectrometer from Edinburgh Instruments, equipped with an R928 photomultiplier or an NIR 301/2 (InGaAs) detector (900–1700 nm, 100 ns response time), and an iCCD camera from Andor. A

144

VI Experimental Part frequency-tripled Quantel Brilliant b laser was used as an excitation source for measuring the Re-complexes. For the Ru-complexes, the output frequency was doubled. The duration of the laser excitation pulse was approximately 10 ns with a repetition frequency of 10 Hz. Unless otherwise stated, transient absorption spectra were obtained as time-averaged signals over a detection period of 200 ns directly after excitation. For all optical measurements Quartz cuvettes from Starna were used.

145

VI Experimental Part

VI.2 Syntheses of the OTA donor OTA–pre Br

Br N O

O

Synthesis according to literature procedure.97 Bis(4-bromophenyl)amine (2.00 g, 6.12 mmol), 4-(N,N-dimethylamino)pyridine (149.5 mg, 1.22 mmol), and di-tert-butyl dicarbonate (2.11 mg, 9.17 mmol) were dissolved in 15 mL dry THF under N2 atmosphere. The yellow solution was heated to reflux for 3 h. After cooling to room temperature and evaporation of the solvent the orange crude solid was filtered over a pluck of SiO2. With an eluent mixture of pentane/DCM 2:1 pure product was obtained as a white solid (2.60 g, 6.09 mmol, 100%). 1 H NMR (400 MHz, CDCl3): 7.41–7.45 (m, 4H, AA’BB’), 7.04–7.08 (m, 4H, AA’BB’), 1.44 (s, 9H, tBu).

OTA–Boc OCH3

OCH3

N

N

H3CO

N O

OCH3 O

Synthesis by adapting a literature procedure.97 OTA–pre (1.00 g, 2.34 mmol), dianisylamine (1.34 g, 5.86 mmol), NaOtBu (4.5 g, 46.8 mmol), Pd(dba)2 (67.2 mg, 0.12 mmol), and (HPtBu3)BF4 (34.2 mg, 0.12 mmol) were dissolved in dry, de-oxygenated toluene (30 mL) under N2 atmosphere. The mixture was stirred at 120 °C for 19 h. 100 mL H2O was added to the cool mixture and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over MgSO4 and evaporation of the solvents the crude product was subjected to column chromatography. With DCM as eluent the pure product was obtained as a beige solid (1.42 g, 1.96 mmol, 84%). 1 H NMR (400 MHz, acetone-d6): 7.05–7.09 (m, 4H, AA’BB’), 7.00–7.04 (m, 8H, AA’BB’),

146

VI Experimental Part 6.87–6.91 (m, 8H, AA’BB’), 6.78–6.82 (m, 4H, AA’BB’), 3.78 (s, 12H, OCH3), 1.41 (s, 9H, Bu).

t

OTA OCH3

OCH3

N

N

H3CO

N H

OCH3

Synthesis according to literature procedure.95 OTA–Boc (2.00 g, 2.76 mmol) was dissolved in 40 mL acetone under N2 atmosphere. TFA (10 mL, 0.13 mol) was added dropwise and stirring at RT was continued over night. The reaction mixture was evaporated to dryness. Column chromatography on basic alumina with EtOAc as eluent yielded the product as a beige solid (1.70 g, 2.73 mmol, 99%). 1 H NMR (400 MHz, acetone-d6 + TFA): 7.37–7.40 (m, 4H, AA’BB’), 7.14–7.18 (m, 8H, AA’BB’), 6.94–6.98 (m, 8H, AA’BB’), 6.83–6.87 (m, 4H, AA’BB’), 3.80 (s, 12H, OCH3).

VI.3 Syntheses of General Building Blocks TAA-ref H3CO

OCH3 N

4-Bromotoluene (895.2 mg, 5.23 mmol), dianisylamine (1.00 g, 4.36 mmol), NaOtBu (8.38 g, 8.72 mol), Pd(dba)2 (125.4 mg, 0.22 mmol), and (HPtBu3)BF4 (63.8 mg, 0.22 mmol) were suspended in dry toluene (50 mL), de-oxygenated and heated to reflux for 23 h. 100 mL H2O was added to the cool reaction mixture and then extracted with DCM (3 x 50 mL). After drying over MgSO4 the crude product was purified by column chromatography with a pentane/DCM (1:1) eluent mixture. Pure product was obtained as a brown sticky solid (1.17 g, 3.66 mmol, 84%). 1 H NMR (400 MHz, CDCl3): 7.07 (td, J = 6.6 Hz, 2.5 Hz, 6H, AA’BB’), 6.97–6.93 (m, 2H, AA’BB’), 6.89–6.82 (m, 4H, AA’BB’), 3.83 (s, 6H, OCH3), 2.34 (s, 3H, CH3).

147

VI Experimental Part

OTA-ref OCH3

OCH3

N

N

H3CO

N

OCH3

Synthesis based on a literature procedure.94 OTA (200.0 mg, 0.32 mmol), bromobenzene (0.05 mL, 0.48 mmol), KOtBu (718.1 g, 6.40 mmol), Pd(dba)2 (18.4 mg, 0.03 mmol), and (HPtBu3)BF4 (9.30 mg, 0.03 mmol) were dissolved in dry, de-oxygenated toluene (15 mL) under N2 atmosphere. The mixture was stirred at 120 °C over night. 100 mL EtOAc was added to the cool mixture and the organic phase was washed with H2O (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With EtOAc as eluent and subsequent recrystallization from MeOH, pure product was obtained as a beige solid (157.0 mg, 0.22 mmol, 70%). 1 H NMR (400 MHz, C6D6): 7.24 (d, J = 7.9 Hz, 2H, ph), 7.14–7.07 (m, 14H, AA’BB’), 7.06–7.01 (m, 4H, AA’BB’ + ph), 6.81 (t, J = 7.2 Hz, 1H, ph), 6.74–6.68 (m, 8H, AA’BB’), 3.30 (s, 12H, OCH3).

Br–bpy

Br N

N

Synthesis according to literature procedure.118 2-pyridylzinc bromide (42.3 mL, 21.1 mmol) and Pd(PPh3)4 were mixed under N2 atmosphere and 5-bromo-2-iodopyridine (5.00 g, 17.6 mmol) in 35 mL dry THF was added drop wise. The mixture was stirred at room temperature for 24 h. 150 mL basic EDTA-solution was added and the product was extracted with DCM (3 x 100 mL). The combined organic phases were washed with 100 mL brine and dried over MgSO4. After evaporation of the solvents the crude product was subjected to column chromatography (SiO2, pentane/Et2O 2:1 + 1 % NEt3) to yield an off-white solid (3.22 g, 13.7 mmol, 78%). 1 H NMR (400 MHz, CDCl3): 8.71–8.72 (dd, J = 2.4 Hz, 0.8 Hz, 1H, bpy), 8.65–8.67 (ddd, J = 4.9 Hz, 1.7 Hz, 0.9 Hz, 1H, bpy), 8.35–8.38 (dt, J = 8.0 Hz, 1.1 Hz, 1H, bpy), 8.30–8.33

148

VI Experimental Part (dd, J = 8.5 Hz, 0.8 Hz, 1H, bpy), 7.91–7.95 (dd, J = 8.5 Hz, 2.4 Hz, 1H, bpy), 7.78–7.84 (td, J = 7.8 Hz, 1.8 Hz, 1H, bpy), 7.29–7.34 (ddd, J = 7.5 Hz, 4.8 Hz, 1.2 Hz, 1H, bpy).

bpy–SnBu3

N

SnBu3

N

Synthesis according to literature procedure.119 Br–bpy (530.0 mg, 2.25 mmol) was dissolved in 15 mL Et2O under N2 atmosphere and cooled to −78 °C in an acetone/dry ice bath. n-BuLi (1.55 mL, 2.48 mmol, 1.6 M) was added drop wise and the dark red mixture was stirred for 1 h. After addition of tributyltin chloride (0.67 mL, 2.48 mmol) the reaction was stirred 1.5 h at −78 °C and over night at RT. Filtration of the solid and evaporation to dryness gave a orange oil, which was purified by column chromatography (SiO2, pentane/EtOAc 90:10 + 1% NEt3) to yield a yellow oil (0.5 g, 1.30 mmol, 58%). 1 H NMR (400 MHz, CDCl3): 8.69 (t, J = 1.3 Hz, 1H, bpy), 8.67 (ddd, J = 4.8 Hz, 1.8 Hz, 0.9 Hz, 1H, bpy), 8.40–8.36 (m, 1H, bpy), 8.32 (dd, J = 7.7 Hz, 1.0 Hz, 1H, bpy), 7.89 (dd, J = 7.7 Hz, 1.6 Hz, 1H, bpy), 7.80 (td, J = 7.7 Hz, 1.8 Hz, 1H, bpy), 7.28 (ddd, J = 7.6 Hz, 4.8 Hz, 1.2 Hz, 1H, bpy), 1.61–1.51 (m, 6H, Bu3), 1.34 (dq, J = 14.5 Hz, 7.3 Hz, 6H, Bu3), 1.15–1.09 (m, 6H, Bu3), 0.88 (t, J = 7.3 Hz, 9H, Bu3).

Br–bpy–Br

Br

Br N

N

Synthesis by adapting a literature procedure.120 5-bromo-2-iodopyridine (10.0 g, 35.2 mmol) was dissolved in 60 mL dry m-xylene and hexa-n-butylditin (8.90 mL, 17.61 mmol) was added to the solution. After thorough degassing of the mixture, Pd(PPh3)4 (814.0 mg, 0.70 mmol) was added and the reaction mixture was de-oxygenated again. The yellow suspension was heated to 180 °C for 3 d. Afterwards, the black solid was removed by hot filtration. Solvent removal and recrystallization from chloroform yielded the product as a beige solid (4.44 g, 14.1 mmol, 80%). 1 H NMR (400 MHz, CDCl3): 8.70 (m, 2H, bpy), 8.27–8.30 (dd, 2H, bpy), 7.92–7.96 (dd, 2H, bpy).

149

VI Experimental Part

TMS–xy–Br

TMS

Br

Synthesis according to literature procedure.119 A solution of 2,5-dibromo-1,4-di-methylbenzene (15.0 g, 57.3 mmol) in dry THF (130 mL) was cooled to −78 °C under a N2 atmosphere. n-BuLi (25.2 mL, 63.0 mmol, 2.5 M) was then added dropwise and the mixture was stirred for 2 h at −78 °C. After addition of trimethylsilyl chloride (8.05 mL, 63.0 mmol) the yellow solution was stirred over night at RT. The solvents were removed in vacuo and the residue was taken up in water (50 mL) and DCM (50 mL). The aqueous phase was extracted with DCM (3 x 50 mL). The combined organic phases were washed with brine (2 x 50 mL) and dried over Na2SO4. The dried crude yellow liquid was purified by filtering over a pluck of SiO2 with pentane as eluent. This gave the product as a colorless liquid (13.8 g, 53.5 mmol, 93%). 1 H NMR (400 MHz, CDCl3): 7.39 (s, 1H, xy), 7.32 (s, 1H, xy), 2.44 (s, 3H, CH3), 2.41 (s, 3H, CH3), 0.36 (s, 9H, Si(CH3)3).

TMS–xy–B(OH)2

TMS

B(OH)2

Synthesis according to literature procedure.121 TMS–xy–Br (5.20 g, 20.2 mmol) was dissolved in 90 mL dry THF under N2 atmosphere and cooled to −78 °C. n-BuLi (9.70 mL, 24.24 mmol, 2.5 M) was added drop wise and the mixture was stirred at room temperature for 1.5 h. After cooling to −78 °C again, trimethyl borate (2.70 mL, 24.24 mmol) was added drop wise and the reaction was stirred over night at RT. The colorless suspension was acidified with 50 mL HCl (6 N) and extracted with DCM (3 x 50 mL). The combined organic phases were dried over MgSO4. Evaporation of the solvents gave a yellowish solid with was purified by column chromatography (SiO2, pentane/Et2O 1:1) to yield pure white product (3.37 g, 15.2 mmol, 75%). 1 H NMR (400 MHz, CDCl3): 7.94 (s, 1H, xy), 7.36 (s, 1H, xy), 2.76, (s, 3H, CH3), 2.50 (s, 3H, CH3), 0.36 (s, 9H, Si(CH3)3).

150

VI Experimental Part

VI.4 Syntheses of ligands and metal complexes with rhenium fac Re(bpy)(CO)3Cl

N

N OC

Cl

Re

OC

CO

Synthesis according to literature procedure.101 Re(CO)5Cl (1.00 g, 2.77 mmol) and 2,2’-bipyridine (433.0 mg, 2.77 mmol) were dissolved in 150 mL toluene and heated to reflux for 1 h. After cooling to room temperature and keeping in fridge over night the yellow solid was filtered, washed with Et2O and dried in vacuum. The yellow product (1.24 g, 2.68 mmol, 97%) was used without further characterization.

fac Re(bpy)(CO)3OTf

N

N OC OC

OTf

Re

CO

Synthesis according to literature procedure.102 fac Re(bpy)(CO)3Cl (357.0 mg, 0.77 mmol) was suspended in 5 mL DCM under N2 atmosphere and trifluoromethanesulfonic acid (0.41 mL, 4.64 mmol) was added dropwise. The clear orange mixture was stirred at RT for 1 h. Upon drop wise addition of Et2O to the stirred reaction mixture a light yellow solid formed which was filtered, washed with Et2O and dried in vacuum (399.0 mg, 0.69 mmol, 90%). This Re-precursor was used without further characterization.

py–Br–py Br

N

N

151

VI Experimental Part 1,3,5-tribromobenzene (0.50 g, 1.59 mmol), Na2CO3 (3.00 g, 28.3 mmol), and 4-pyridine boronic acid (586.3 mg, 4.77 mmol) were dissolved in 10 mL water and 20 mL THF under N2 atmosphere and the suspension was de-oxygenated for 30 min. After addition of Pd(PPh3)4 (184.0 mg, 0.16 mmol) and additional degassing the mixture was heated to 90 °C for 24 h. The cool solution was mixed with 50 mL water and extracted with DCM (3 x 50 mL). The combined organic phases were dried over Na2SO4, the solvents were removed and subsequent column chromatography (SiO2, DCM + 2% NEt3) gave a light green solid (0.18 g, 0.58 mmol, 36%). 1 H NMR (400 MHz, CDCl3): 8.75–8.69 (m, 4H, py), 7.83 (d, J = 1.6 Hz, 2H), 7.77 (t, J = 1.6 Hz, 1H), 7.54–7.49 (m, 4H, py).

py–TAA–py H3CO

OCH3 N

N

N

Py–Br–py (66.0 mg, 0.21 mmol), dianisylamine (58.3 mg, 0.25 mmol), NaOtBu (404.0 mg, 4.20 mmol), Pd(dba)2 (6.00 mg, 0.01 mmol), and (HPtBu3)BF4 (3.00 mg, 0.01 mmol) were dissolved in dry, de-oxygenated toluene (5 mL) under N2 atmosphere. The mixture was stirred at 100 °C for 22.5 h. 50 mL H2O was added to the cool mixture and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With DCM and 1% NEt3 as eluent mixture the pure product was obtained as a beige solid (80.0 mg, 0.17 mmol, 83%). 1 H NMR (400 MHz, CDCl3): 8.62 (m, 4H, py), 7.41 (m, 4H, py), 7.29 (t, J = 1.6 Hz, 1H), 7.20 (d, J = 1.6 Hz, 2H), 7.16–7.12 (m, 4H, AA’BB’), 6.90–6.86 (m, 4H, AA’BB’), 3.82 (s, 6H, OCH3).

py–OTA–py OCH3

OCH3

N

N

H3CO

N

N

152

OCH3

N

VI Experimental Part Py–Br–py (180.5 mg, 0.58 mmol) and OTA (0.30 g, 0.48 mmol), together with NaOtBu (1.11 g, 11.6 mmol), Pd(dba)2 (13.8 mg, 0.02 mmol), and (HPtBu3)BF4 (6.90 mg, 0.02 mmol) were dissolved in dry, de-oxygenated toluene (25 mL) under N2 atmosphere. The mixture was stirred at 125 °C for 19 h. 100 mL H2O was added to the cool mixture and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over MgSO4 and evaporation of the solvents the crude product was subjected to column chromatography. With DCM and 2% NEt3 as eluent mixture and subsequent recrystallization (hexane) the pure product was obtained as an olive-green solid (292.0 mg, 0.34 mmol, 71%). 1 H NMR (400 MHz, CDCl3): 8.66 (d, J = 4.7 Hz, 4H, py), 7.44 (d, J =6.2 Hz, 4H, py), 7.29 (s, 3H), 7.08–7.04 (m, 8H, AA’BB’), 7.02–6.99 (m, 4H, AA’BB’), 6.91–6.87 (m, 4H, AA’BB’), 6.85–6.81 (m, 8H, AA’BB’), 3.79 (s, 12H, OCH3).

Re2–TAA

H3CO

OCH3

2+ 2NO3-

N

CO OC

CO N

N

Re OC

N N

CO Re

N

CO N

Py–TAA–py (100.0 mg, 0.22 mmol) and Re(bpy)(CO)3OTf (275.4 mg, 0.48 mmol) were dissolved in 10 mL MeOH. The mixture was de-oxygenated for 45 min and heated to 85 °C over night. After evaporation to dryness the crude product was subjected to column chromatography (SiO2, acetone → acetone/H2O/sat. KNO3 200:9:1), the orange fraction was collected and the solvents were removed again. 50 mL H2O was added and the product was extracted with DCM (3 x 50 mL), dried over Na2SO4 and the solvents were removed in vacuum. The solid was recrystallized from hexane yielding an orange solid (160.7 mg, 0.11 mmol, 50%). 1 H NMR (400 MHz, CD3CN): 9.21 (ddd, J = 5.5 Hz, 1.6 Hz, 0.8 Hz, 4H, bpy), 8.37 (dt, J = 8.3 Hz, 1.0 Hz, 4H, bpy), 8.25 (td, J = 7.9 Hz, 1.5 Hz, 4H, bpy), 8.20–8.15 (m, 4H, py), 7.77 (ddd, J = 7.6 Hz, 5.5 Hz, 1.3 Hz, 4H, bpy), 7.37–7.32 (m, 4H, py), 7.19 (t, J = 1.6 Hz, 1H), 7.07–7.02 (m, 4H, AA’BB’), 6.97 (d, J = 1.6 Hz, 2H), 6.89–6.84 (m, 4H, AA’BB’), 3.75 (s, 6H, OCH3). ESI-MS calculated (m/z) for C56H41N7O8Re2: 656.6061; found: 656.6075. Elemental analysis calculated (%) for C56H41N9O14Re2 · 2 H2O: C 45.68, H 3.08, N 8.56; found: C 45.62, H 3.21, N 8.63.

153

VI Experimental Part

Re2–OTA OCH3

OCH3

N

N

H3CO

2+ 2NO3-

N

OCH3

CO OC

CO N

N

Re OC

CO Re

N

N

N

CO N

Py–OTA–py (100.0 mg, 0.12 mmol) and Re(bpy)(CO)3OTf (148.2 mg, 0.26 mmol) were dissolved in 10 mL methanol and 3 mL chloroform. The orange solution was deoxygenated for 25 min and heated to 85 °C for 1.5 d. After evaporation to dryness the crude product was subjected to column chromatography (SiO2, acetone → acetone/ H2O/sat. KNO3 200:9:1). The obtained orange solid was dissolved in 50 mL H2O and extracted with DCM (3 x 50 mL), dried over Na2SO4 and the solvents were removed in vacuum. The solid was recrystallized from hexane yielding a brown solid (83.0 mg, 0.04 mmol, 37%). 1 H NMR (400 MHz, CD3CN): 9.23 (ddd, J = 5.5 Hz, 1.6 Hz, 0.7 Hz, 4H, bpy), 8.40 (dt, J = 8.3 Hz, 1.1 Hz, 4H, bpy), 8.27 (td, J = 7.9 Hz, 1.6 Hz, 4H, bpy), 8.23–8.19 (m, 4H, py), 7.80 (ddd, J = 7.7 Hz, 5.5 Hz, 1.3 Hz, 4H, bpy), 7.41–7.36 (m, 4H, py), 7.20 (bs, 1H), 7.12 (bs, 2H), 6.88 (td, 24H, AA’BB’), 3.78 (s, 12H, OCH3). ESI-MS calculated (m/z) for C82H63N9O10Re2: 853.6902; found: 853.6915. Elemental analysis calculated (%) for C82H63N11O16Re2 · 3 H2O: C 52.25, H 3.69, N 8.17; found: C 52.14, H 3.79, N 8.09.

AQ–xy–TMS

O TMS O

Synthesis followed a previously published protocol of ours.88 2-bromoanthraquinone (3.32 g, 11.6 mmol), TMS–xy–B(OH)2 (3.08 g, 13.9 mmol), and Na2CO3 (3.67 g, 34.7 mmol) were dissolved in 12 mL water, 10 mL EtOH and 60 mL toluene under N2 atmosphere and the suspension was de-oxygenated for 15 min. After addition of Pd(PPh3)4 (1.33 g, 1.16 mmol) and additional degassing the mixture was

154

VI Experimental Part refluxed for 1.5 d. The reaction solution was mixed with 100 mL H2O and extracted with DCM (3 x 50 mL). The combined organic phases were dried over Na2SO4, the solvents were removed and subsequent column chromatography (SiO2, pentane/DCM 1:1) gave a yellow solid (4.44 g, 11.6 mmol, quant.). 1 H NMR (400 MHz, CDCl3): 8.32–8.38 (m, 3H, AQ), 8.30 (d, J = 1.8 Hz, 1H, AQ), 7.82 (m, 2H, AQ), 7.76–7.79 (dd, J = 8.0 Hz, 1.9 Hz, 1H, AQ), 7.40 (s, 1H, xy), 7.11 (s, 1H, xy), 2.49 (s, 3H, CH3), 2.30 (s, 3H, CH3), 0.38 (s, 9H, Si(CH3)3).

AQ–xy–I

O I O

A previously published protocol was followed.88 AQ–xy–TMS (1.97 g, 5.12 mmol) was dissolved in 10 mL DCM under N2 atmosphere and cooled to 0 °C. A solution of ICl (563 µL, 10.8 mmol) in 40 mL MeCN was added very slowly, the yellow suspension was stirred at 0 °C for 15 min and at RT over night. After quenching with 200 mL Na2S2O3-solution (5% in H2O) and extraction with DCM (3 x 50 mL) the combined organic phases were washed with 50 mL H2O and dried over Na2SO4. Evaporation to complete dryness yielded a yellow solid (2.2 g, 5.12 mmol, quant.). 1 H NMR (400 MHz, CDCl3): 8.31–8.37 (m, 3H, AQ), 8.25 (d, J = 1.8 Hz, 1H, AQ), 7.80– 7.85 (m, 2H, AQ), 7.79 (s, 1H, xy), 7.72–7.74 (dd, J = 8.0 Hz, 1.8 Hz, 1H, AQ), 7.14 (s, 1H, xy), 2.45 (s, 3H, CH3), 2.23 (s, 3H, CH3).

AQ–xy–bpy

O N

N

O

Preparation by adapting a previously published protocol.88 AQ–xy–I (820.3 mg, 1.87 mmol) and bpy–SnBu3 (1.00 g, 2.25 mmol) were dissolved in 80 mL m-xylene under N2 atmosphere and de-oxygenated for 30 min. After addition of Pd(PPh3)4 (108.0 mg, 0.09 mmol) the reaction mixture was de-oxygenated again by 3 cycles of vacuum/N2 and heated to 150 °C for 70 h. 150 mL H2O was added to the cool, black suspension and extracted with DCM (3 x 50 mL). The combined

155

VI Experimental Part organic phases were dried over MgSO4, the solvents were evaporated and the crude product was subjected to column chromatography (SiO2, DCM, then DCM+1% NEt3). Recrystallization from pentane gave a yellow solid (640.0 mg, 1.37 mmol, 73%). 1 H NMR (400 MHz, CDCl3): 8.73 (ddd, J = 5.5 Hz, 2.0 Hz, 0.9 Hz, 2H, bpy), 8.50 (dd, J = 8.1 Hz, 0.8 Hz, 1H, bpy), 8.46 (dt, J = 8.0 Hz, 1.1 Hz, 1H, bpy), 8.40 (d, J = 8.0 Hz, 1H, bpy), 8.38–8.32 (m, 3H, AQ), 7.89–7.80 (m, 6H, 4 AQ + 1 bpy + 1 xy), 7.34 (ddd, J = 7.5 Hz, 4.8 Hz, 1.8 Hz, 1H, bpy), 7.27 (s, 1H, xy), 2.36 (d, J = 3.2 Hz, 6H, CH3).

py–xy–TMS

TMS

N

TMS–xy–B(OH)2 (1.37 g, 6.17 mmol) and 4-bromopyridine hydrochloride (1.00 g, 5.14 mmol), together with Na2CO3 (3.26 g, 30.8 mmol) were suspended in a mixture of H2O/EtOH/toluene (8/6/20; v:v:v) under N2 atmosphere and de-oxygenated for 20 min. After addition of Pd(PPh3)4 (59.4 mg, 0.05 mmol) the reaction mixture was deoxygenated again for 10 min and heated to reflux for 1.5 d. After addition of 100 mL H2O to the cool reaction mixture the product was extracted with DCM (3 x 50 mL), dried over MgSO4 and the solvents were evaporated. Column chromatography on SiO2 with an eluent mixture of pentane/DCM 2/1 + 1% NEt3) gave the product as a light yellow solid (1.30 g, 5.09 mmol, 99%). 1 H NMR (400 MHz, CDCl3): 8.65–8.63 (m, 2H, py), 7.37 (s, 1H, xy), 7.27–7.26 (m, 2H, py), 7.02 (s, 1H, xy), 2.46 (s, 3H, CH3), 2.26 (s, 3H, CH3), 0.36 (s, 9H, Si(CH3)3).

py–xy–I

N

I

Py–xy–TMS (1.30 g, 5.09 mmol) was dissolved in 5 mL DCM under N2 atmosphere and cooled to 0 °C. At this temperature a solution of ICl (520 µL, 10.2 mmol) in 20 mL MeCN was added dropwise and the resulting suspension was stirred at RT over night. After quenching with 200 mL sat. Na2S2O3-solution the product was extracted with DCM (3 x 50 mL), dried over Na2SO4 and the solvents were evaporated. The crude product was filtered over a pluck of SiO2 with an eluent mixture of pentane/DCM

156

VI Experimental Part 2/1 + 1% NEt3) and gave pure product as beige solid (1.52 g, 4.92 mmol, 97%). 1 H NMR (400 MHz, CDCl3): 8.68–8.61 (m, 2H, py), 7.76 (s, 1H, xy), 7.25–7.18 (m, 2H), 7.06 (s, 1H, xy), 2.43 (s, 3H, CH3), 2.19 (s, 3H, CH3).

py–xy–OTA H3CO

N

OCH3

N

OCH3

N

N

H3CO

Py–xy–I (177.2 mg, 0.57 mmol) and OTA (429.0 mg, 0.69 mmol), together with KOtBu (1.28 g, 11.4 mmol), Pd(dba)2 (32.8 mg, 0.06 mmol), and (HPtBu3)BF4 (16.5 mg, 0.06 mmol) were dissolved in dry, de-oxygenated toluene (15 mL) under N2 atmosphere. The mixture was heated to 100 °C for 2.5 d. 150 mL H2O was added and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. First with pure DCM, then with DCM + 1% NEt3 as eluents the pure product was obtained as a brown solid (0.43 g, 0.53 mmol, 94%). 1 H NMR (400 MHz, CDCl3): 8.64–8.61 (m, 2H, py), 7.29–7.26 (m, 2H, py), 7.06–6.99 (m, 10H, xy + AA’BB’), 6.87–6.77 (m, 16H, AA’BB’), 3.78 (s, 12H, OCH3), 2.20 (s, 3H, CH3), 2.05 (s, 3H, CH3).

AQ–Re–Cl

O N

N O

OC OC

Re

Cl CO

AQ–xy–bpy (150.0 mg, 0.32 mmol) and Re(CO)5Cl (116.3 mg, 0.32 mmol) were dissolved in 25 mL toluene and heated to reflux over night. After cooling to RT the yellow solid was filtered, washed with Et2O and dried in vacuum (224.5 mg, 0.29 mmol, 91%).

157

VI Experimental Part 1

H NMR (400 MHz, CDCl3): 9.13–9.07 (m, 2H, bpy), 8.42 (d, J = 8.0 Hz, 1H, bpy), 8.38– 8.32 (m, 3H, AQ), 8.26 (dd, J = 9.7 Hz, 8.0 Hz, 2H, AQ + bpy), 8.11 (ddd, J = 8.6 Hz, 6.9 Hz, 1.9 Hz, 2H, bpy), 7.87–7.80 (m, 3H, AQ), 7.57 (ddd, J = 7.6 Hz, 5.5 Hz, 1.2 Hz, 1H, bpy), 7.30 (d, J = 6.6 Hz, 2H, xy), 2.39 (d, J = 11.0 Hz, 6H, CH3).

AQ–Re–OTf

O N

N OC

O

OTf

Re

OC

CO

AQ–Re–Cl (201.0 mg, 0.26 mmol) was suspended in 15 mL dry DCM. Trifluoromethanesulfonic acid (1.50 mL, 17 mmol) was added dropwise at RT and the solution was stirred for 2 h. Et2O was added slowly the reaction mixture and the product precipitated as a yellow solid. After keeping in the fridge over night the solid was filtered, washed with Et2O and dried in vacuum (190.3 mg, 0.21 mmol, 83%). 1 H NMR (400 MHz, CDCl3): 9.16–9.10 (m, 2H, bpy), 8.42 (dd, J = 8.0 Hz, 2.0 Hz, 1H, bpy), 8.39–8.32 (m, 3H, AQ), 8.29 (t, J = 7.6 Hz, 2H, AQ + bpy), 8.24–8.15 (m, 2H, bpy), 7.8–7.79 (m, 3H, AQ), 7.66 (ddd, J = 7.6 Hz, 5.4 Hz, 1.3 Hz, 1H, bpy), 7.30 (d, J = 12.6 Hz, 2H, xy), 2.42–2.35 (m, 6H, CH3).

AQ–Re–OTA 2+ 2 OTf -

H3CO

N

O N

N O

OC OC

OCH3

Re

N

N

CO OCH3

H3CO

AQ–Re–OTf (80.7 mg, 0.08 mmol) and py–xy–OTA (74.0 mg, 0.10 mmol) were suspended in 5 mL MeOH and 5 mL CHCl3, de-oxygenated for 20 min and heated to reflux for 43 h. The solvents were evaporated and the residue was subjected to column chromatography with DCM + 1% MeOH, then 3% MeOH, as eluents. The obtained solid was dissolved in a minimal amount of DCM and precipitated in Et2O

158

VI Experimental Part to get pure product as a dark orange solid (64.0 mg, 0.04 mmol, 47%). 1 H NMR (400 MHz, CD3CN): 9.27–9.23 (m, 1H), 9.19 (d, J = 1.7 Hz, 1H), 8.44 (dd, J = 8.2 Hz, 4.2 Hz, 2H), 8.35–8.32 (m, 1H), 8.31–8.24 (m, 7H), 7.88 (dt, J = 6.8 Hz, 1.5 Hz, 3H), 7.83–7.77 (m, 1H), 7.44 (s, 1H), 7.37 (s, 1H), 7.24 (s, 2H) 6.86 (bs, 26 H), 3.73 (s, 12H), 2.35 (d, J = 2.7 Hz, 12H). ESI-MS calculated (m/z) for C88H70N6O9Re: 1541.4765; found: 1541.4774. Elemental analysis calculated (%) for C89H70F3N6O12ReS · Et2O: C 63.29, H 4.57, N 4.76; found: C 63.01, H 4.68, N 5.11.

Re–ref + OTf N

N OC OC

Re

N CO

fac Re(bpy)(CO)3OTf (0.20 g, 0.35 mmol) was dissolved in 20 mL MeOH and pyridine (0.17 mL, 2.09 mmol) was added under N2 atmosphere. The yellow reaction mixture was heated to 85 °C for 24 h. After cooling to RT the solvents were evaporated, the residue was taken up in little DCM and precipitated in hexane. This afforded the product as a yellow solid (193.5 mg, 0.30 mmol, 85%). 1 H NMR (400 MHz, CD3CN): 9.22 (ddd, J = 5.5 Hz, 1.5 Hz, 0.8 Hz, 2H, bpy), 8.37 (dt, J = 8.3 Hz, 1.0 Hz, 2H, bpy), 8.30–8.20 (m, 4H, bpy + py), 7.85 (tt, J = 7.7 Hz, 1.6 Hz, 1H, py), 7.79 (ddd, J = 7.6 Hz, 5.5 Hz, 1.3 Hz, 2H, bpy), 7.33–7.26 (m, 2H, py).

VI.5 Syntheses of ligands and metal complexes with ruthenium Ru–pre

N

N N

Cl

Ru N

Cl

RuCl3 · 3 H2O (1.98 g, 9.55 mmol), 2,2’-bipyridine (3.00 g, 19.2 mmol), and lithium chloride (2.83 g, 66.9 mmol) were suspended in 17 mL dry DMF under N2 atmosphere and

159

VI Experimental Part heated to reflux for 16 h. The cooled reaction mixture was concentrated, acetone was added and the product was crystallized in the fridge over night. The precipitate was filtered, washed with acetone, H2O, and Et2O. After drying in vacuum the product was obtained as a dark violet solid (673.0 mg, 1.39 mmol, 15%). This Ru-precursor was used without further characterization.

AQ–xy–B–ester

O

O B O O

AQ–xy–I (200.0 mg, 0.46 mmol), bis-pinacolato diboron (173.8 mg, 0.68 mmol), and KOAc (198.6 mg, 2.02 mmol) were suspended in dry DMF (10 mL) and de-oxygenated for 30 min. After addition of PdCl2(PPh3)2 (8.10 mg, 0.01 mmol) the mixture was heated to 100 °C for 24 h. The DMF was evaporated and the residue was taken up in 50 mL DCM and washed with H2O repeatedly. After drying over Na2SO4 and evaporation of the solvents, the crude product was subjected to column chromatography. With pentane/DCM 1:1 as eluent the pure product was obtained as yellow crystals (183.0 mg, 0.42 mmol, 91%). 1 H NMR (400 MHz, CDCl3): 8.38–8.30 (m, 3H, AQ), 8.28 (dd, J = 1.8 Hz, 0.5 Hz, 1H, AQ), 7.85–7.78 (m, 2H, AQ), 7.76 (dd, J = 8.0 Hz, 1.8 Hz, 1H, AQ), 7.72 (s, 1H, xy), 7.11 (s, 1H, xy), 2.56 (s, 3H, CH3), 2.28 (s, 3H, CH3), 1.38 (s, 12H, CH3).

AQ–xy–bpy–Br O Br N

N

O

AQ–xy–B–ester (200.0 mg, 0.46 mmol), Br–bpy–Br (286.5 mg, 0.91 mmol), and Na2CO3 (146.3 mg, 1.38 mmol) were suspended in 10 mL THF and 3 mL H2O under N2 atmosphere and de-oxygenated for 20 min. After addition of Pd(PPh3)4 (35.0 mg, 0.03 mmol) the reaction mixture was de-oxygenated again for 10 min and heated to reflux over night. After cooling the reaction mixture to RT, the product was filtered and washed with MeOH. Column chromatography on SiO2 with an eluent mixture of DCM + 1% NEt3 yielded the product as a yellow solid (0.23 mg, 0.42 mmol, 92%). 1 H NMR (400 MHz, CDCl3): 8.76 (dd, J = 3.1 Hz, 2.4 Hz, 1H, bpy), 8.71 (dd, J = 2.2 Hz,

160

VI Experimental Part 0.9 Hz, 1H, bpy), 8.47 (dd, J = 8.1 Hz, 0.9 Hz, 1H, bpy), 8.43–8.38 (m, 2H, bpy), 8.378.34 (m, 3H, AQ), 7.98 (dd, J = 8.5 Hz, 2.4 Hz, 1H, bpy), 7.85–7.82 (m, 4H, AQ), 7.27 (s, 1H, xy), 7.25 (s, 1H, xy), 2.36 (s, 3H, CH3), 2.35 (s, 3H, CH3).

AQ–xy–bpy–xy–TMS O TMS N

N

O

AQ–xy–bpy–Br (196.0 mg, 0.36 mmol), TMS–xy–B(OH)2 (95.8 mg, 0.43 mmol), and Na2CO3 (228.9 mg, 2.16 mmol) were suspended in 4 mL H2O, 3 mL EtOH and 10 mL toluene under N2 atmosphere and de-oxygenated for 20 min. Following the addition of Pd(PPh3)4 (41.6 mg, 0.04 mmol) the reaction mixture was de-oxygenated again for 15 min and heated to reflux for 2 d. After addition of 100 mL H2O the product was extracted with DCM (3 x 50 mL), dried over Na2SO4 and the solvents were evaporated. Column chromatography on SiO2 with an eluent mixture of DCM + 1% MeOH and subsequent recrystallization from pentane gave the product as a light yellow solid (0.17 mg, 0.26 mmol, 73%). 1 H NMR (400 MHz, CDCl3): 8.74 (dd, J = 14.3 Hz, 2.2 Hz, 2H, bpy), 8.52 (t, J = 8.2 Hz, 2H, bpy), 8.40 (d, J = 8.0 Hz, 1H, bpy), 8.38–8.33 (m, 3H, AQ), 7.88 (dd, J = 8.1 Hz, 2.2 Hz, 1H, bpy), 7.85–7.80 (m, 4H, AQ), 7.41 (s, 1H, xy), 7.28 (s, 2H, xy), 7.12 (s, 1H, xy), 2.50 (s, 3H, CH3), 2.38 (s, 3H, CH3), 2.36 (s, 3H, CH3), 2.33 (s, 3H, CH3), 0.38 (s, 9H, Si(CH3)3).

AQ–xy–bpy–xy–I O I N

N

O

AQ–xy–bpy–xy–TMS (0.17 g, 0.26 mmol) was dissolved in 6 mL DCM under N2 atmosphere and cooled to 0 °C. At this temperature a solution of ICl (30 µL, 0.56 mmol) in 3 mL MeCN was added dropwise and the resulting suspension was stirred at RT over night. After quenching with 200 mL 5% aqueous Na2S2O3-solution the product was extracted with DCM (3 x 50 mL), dried over Na2SO4 and the solvents were evaporated. The crude product was passed through a pluck of SiO2 with an eluent mixture of DCM + 1% MeOH, which gave pure product as a yellow solid (0.12 g, 0.17 mmol,

161

VI Experimental Part 66%). 1 H NMR (400 MHz, CDCl3): 8.77 (d, J = 1.7 Hz, 1H, bpy), 8.68 (d, J = 1.7 Hz, 1H, bpy), 8.58 (bs, 2H, bpy), 8.41 (dd, J = 7.9 Hz, 0.5 Hz, 1H, bpy), 8.38–8.34 (m, 3H, AQ), 7.93 (d, J = 8.1 Hz, 1H, bpy), 7.87–7.81 (m, 4H, AQ), 7.81 (s, 1H, xy), 7.29–7.27 (m, 2H, xy), 7.16 (s, 1H, xy), 2.47 (s, 3H, CH3), 2.38 (s, 3H, CH3), 2.36 (s, 3H, CH3), 2.27 (s, 3H, CH3).

AQ–bpy–OTA

H3CO

N

OCH3

N

OCH3

O N N

N

O

H3CO

AQ–xy–bpy–xy–I (70.0 mg, 0.10 mmol) and OTA (68.9 mg, 0.11 mmol) were suspended in 10 mL dry, de-oxygenated toluene, together with NaOtBu (192.2 mg, 2.00 mmol), Pd(dba)2 (5.80 mg, 0.01 mmol), and (HPtBu3)BF4 (2.90 mg, 0.01 mmol). The mixture was heated to reflux for 27 h. After evaporation of the solvent the residue was taken up in DCM (50 mL), washed with H2O and dried over Na2SO4. The crude product was subjected to column chromatography with pure DCM then DCM + 1% MeOH as eluents and subsequently recrystallized from hexane. This procedure gave the desired ligand as a brown solid (0.06 g, 0.05 mmol, 50%). 1 H NMR (400 MHz, C6D6): 9.01–8.95 (m, 2H, bpy), 8.95–8.90 (m, 2H, bpy), 8.50 (d, J = 1.8 Hz, 1H, bpy), 8.36 (d, J = 8.0 Hz, 1H, AQ), 8.33–8.25 (m, 2H, AQ), 7.53 (dt, J = 8.1 Hz, 2.2 Hz, 2H, AQ), 7.35 (dd, J = 7.9 Hz, 1.8 Hz, 2H, AQ), 7.27 (s, 1H, bpy), 7.21 (s, 1H, xy), 7.14–7.08 (m, 16H, AA’BB’), 6.99 (d, J = 3.8 Hz, 2H, xy), 6.96 (s, 1H, xy), 6.73 (d, J = 8.7 Hz, 8H, AA’BB’), 3.30 (s, 12H, OCH3), 2.20 (s, 3H, CH3), 2.08 (s, 6H, CH3), 1.99 (s, 3H, CH3).

162

VI Experimental Part

AQ–Ru–OTA 2+ 2PF6-

H3CO

N

OCH3

N

OCH3

O N N O

N Ru

N N

N N H3CO

AQ–bpy–OTA (60.0 mg, 0.05 mmol) and Ru–pre (24.4 mg, 0.05 mmol) were suspended in a de-oxygenated CHCl3/EtOH (3 mL/10 mL) solvent mixture under nitrogen atmosphere and heated to reflux for 2 d. Upon complete consumption of the starting materials, the reaction mixture was evaporated to dryness and the residue was subjected to column chromatography (SiO2, acetone → acetone/H2O/sat. aqueous KNO3 90:10:1). After almost complete evaporation of the acetone the desired complex was precipitated by drop wise addition of sat. aqueous KP6 solution. Filtration and washing with H2O and Et2O gave a dark brown solid (64.0 mg, 0.03 mmol, 68%). 1H NMR (400 MHz, acetone-d6): 8.97 (dd, J = 10.9 Hz, 8.5 Hz, 2H, bpy), 8.85 (td, J = 8.0 Hz, 3.0 Hz, 4H, bpy), 8.39–8.12 (m, 15H, AQ + bpy), 8.06 (t, J = 1.6 Hz, 2H, bpy), 8.01–7.95 (m, 2H, bpy), 7.91 (dd, J = 8.0 Hz, 1.9 Hz, 1H, xy), 7.66 (ddt, J = 7.5 Hz, 5.7 Hz, 1.5 Hz, 2H, bpy), 7.58 (dddd, J = 10.1 Hz, 7.4 Hz, 5.7 Hz, 1.2 Hz, 2H, bpy), 7.24 (d, J = 15.3 Hz, 2H, xy), 7.13 (s, 1H, xy), 6.99 (d, J = 8.1 Hz, 8H, AA’BB’), 6.93–6.68 (m, 16H, AA’BB’), 3.77 (s, 12H, OCH3), 2.29 (s, 3H, CH3), 2.09 (s, 3H, CH3), 2.01 (s, 3H, CH3), 1.97 (s, 3H, CH3). ESI-MS calculated (m/z) for C100H81N9O6Ru: 802.7675; found: 802.7670. Elemental analysis calculated (%) for C100H81F12N9O6P2Ru · 5.5 H2O: C 60.21, H 4.65, N 6.32; found: C 60.30, H 4.52, N 6.15.

bpy–xy–TMS

N

N

TMS

Bpy–Br (1.00 g, 4.25 mmol), TMS–xy–B(OH)2 (1.13 g, 5.10 mmol), and Na2CO3 (1.35 g, 12.8 mmol) were suspended in 8 mL H2O, 20 mL EtOH and 160 mL toluene under N2 atmosphere and de-oxygenated for 1 h. After addition of Pd(PPh3)4 (246.0 mg,

163

VI Experimental Part 0.21 mmol) the reaction mixture was de-oxygenated again for 30 min. The mixture was heated to 90 °C for 1.5 d. 100 mL H2O was added to the cooled reaction mixture and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With pentane/Et2O (3:1) + 1% NEt3 as eluent mixture the product was obtained as a beige solid (1.40 g, 4.21 mmol, 99%). 1 H NMR (400 MHz, CDCl3): 8.72 (d, J = 4.8 Hz, 2H, bpy), 8.48 (t, J = 8.3 Hz, 2H, bpy), 7.83 (td, J = 8.1 Hz, 7.5 Hz, 2.0 Hz, 2H, bpy), 7.43 (s, 1H, xy), 7.31 (ddd, J = 7.4 Hz, 4.8 Hz, 1.1 Hz, 1H, bpy), 7.12 (s, 1H, xy), 2.51 (s, 3H, CH3), 2.34 (s, 3H, CH3), 0.40 (s, 9H, Si(CH3)3).

bpy–xy–I

N

I

N

Bpy–Br (500.0 mg, 1.50 mmol) was dissolved in 3 mL DCM under N2 and cooled to 0°C. ICl (0.17 mL, 3.00 mmol) in 10 mL DCM were added dropwise to the cooled reaction mixture, and it was stirred at RT over night. Excess ICl was quenched by addition of 100 mL sat. aqueous Na2S2O3-solution and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents, the crude product was subjected to column chromatography. With pentane/Et2O (1:1) as eluent mixture the product was obtained as a light yellow solid (0.56 g, 1.45 mmol, 97%). 1 H NMR (400 MHz, acetone-d6): 8.70 (ddd, J = 4.8 Hz, 1.8 Hz, 0.9 Hz, 1H, bpy), 8.65 (dd, J = 2.3 Hz, 0.8 Hz, 1H, bpy), 8.56 (dd, J = 8.2 Hz, 0.8 Hz, 1H, bpy), 8.53 (dt, J = 8.0 Hz, 1.0 Hz, 1H, bpy), 7.98–7.90 (m, 2H, bpy), 7.84 (s, 1H, xy), 7.44 (ddd, J = 7.5 Hz, 4.8 Hz, 1.2 Hz, 1H, bpy), 7.28 (s, 1H, xy), 2.44 (s, 3H, CH3), 2.28 (s, 3H, CH3).

bpy–xy–OTA H3CO

N

N

OCH3

N

OCH3

N

H3CO

164

N

VI Experimental Part Bpy–xy–I (148.6 mg, 0.38 mmol), OTA (200.0 mg, 0.32 mmol), NaOtBu (615.0 mg, 6.40 mmol), Pd(dba)2 (9.20 mg, 0.02 mmol), and (HPtBu3)BF4 (4.60 mg, 0.02 mmol) were dissolved in dry, de-oxygenated toluene (15 mL) under N2 atmosphere. The mixture was heated to 110 °C for 24 h. 150 mL H2O was added to the cooled reaction mixture and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents, the crude product was subjected to column chromatography. With DCM + 1% NEt3 the brown product fraction was eluted and recrystallization from pentane yielded a brown solid (0.28 g, 0.32 mmol, 99%). 1 H NMR (400 MHz, acetone-d6): 8.71–8.67 (m, 2H, bpy), 8.54 (td, J = 8.2 Hz, 0.9 Hz, 2H, bpy), 7.97–7.91 (m, 2H, bpy), 7.42 (ddd, J = 7.5 Hz, 4.8 Hz, 1.2 Hz, 1H, bpy), 7.24 (s, 1H, xy), 7.12 (s, 1H, xy), 7.02–6.98 (m, 8H, AA’BB’), 6.86 (td, J = 5.8 Hz, 5.2 Hz, 2.9 Hz, 16H, AA’BB’), 3.77 (s, 12H, OCH3), 2.27 (s, 3H, CH3), 2.11 (s, 3H, CH3).

Ru–OTA

2+ 2PF6-

H3CO

N

N

OCH3

N

OCH3

N

N Ru

N

N

N

N H3CO

Bpy–xy–OTA (100.0 mg, 0.11 mmol) and Ru–pre (54.9 mg, 0.11 mmol) were suspended in 20 mL EtOH and 6 mL CHCl3 under N2 atmosphere and de-oxygenated for 20 min. After refluxing for 21 h, the solvents were evaporated and the residue was subjected to column chromatography. The product was eluted with a mixture of acetone/H2O 9:1 + 1% sat. aqueous KNO3, the acetone was evaporated and the product precipitated by dropwise addition of a sat. aqueous KPF6 solution. Filtration and drying gave a dark red solid (0.11 g, 0.07 mmol, 64%). 1 H NMR (400 MHz, CD3CN): 8.59–8.47 (m, 6H, bpy), 8.06 (dddt, J = 8.5 Hz, 7.0 Hz, 3.3 Hz, 1.7 Hz, 5H, bpy), 7.98 (td, J = 8.0 Hz, 1.4 Hz, 1H, bpy), 7.88–7.84 (m, 1H, bpy), 7.80–7.76 (m, 1H, bpy), 7.76–7.72 (m, 3H, bpy), 7.66 (d, J = 1.7 Hz, 1H, bpy), 7.45–7.37 (m, 4H, bpy), 7.37–7.31 (m, 1H, bpy), 7.03 (s, 1H, xy), 6.98–6.93 (m, 8H, AA’BB’), 6.93 (s, 1H, xy), 6.86–6.81 (m, 8H, AA’BB’), 6.78–6.69 (m, 8H, AA’BB’), 3.74 (s, 12H, OCH3), 1.97 (s, 3H, CH3), 1.85 (s, 3H, CH3). ESI-MS calculated (m/z) for C78H67N9O4Ru:

165

VI Experimental Part 647.7174; found: 647.7190. Elemental analysis calculated (%) for C78H67F12N9O4P2Ru · 2 CH3COCH3: C 59.29, H 4.68, N 7.41; found: C 59.24, H 4.85, N 7.63.

Br–TMS–Br TMS Br

Br

1,3,5-tribromobenzene (5.00 g, 15.9 mmol) was dissolved in 100 mL Et2O under N2 atmosphere and cooled to -78 °C. After dropwise addition of n-BuLi (6.99 mL, 17.5 mmol) the suspension was stirred at -78 °C for 3 h and TMSCl was added dropwise. The reaction mixture was stirred at RT over night, quenched with 100 mL H2O and extracted with Et2O. The organic phase was dried over Na2SO4, the solvent was evaporated and the crude product was subjected to column chromatography. With pure pentane as eluent the product was obtained as a light yellow liquid (4.34 g, 14.1 mmol, 89%). 1 H NMR (400 MHz, CDCl3): 7.65 (t, J = 1.8 Hz, 1H, ph), 7.52 (d, J = 1.8 Hz, 2H, ph), 0.28 (s, 9H, Si(CH3)3).

B–hex–B

O B O

O B O

1,4-dihexyl-2,5-diiodobenzene (3.53 g, 7.09 mmol), bis-pinacolato diboron (5.40 g, 21.3 mmol), and KOAc (5.57 g, 56.7 mmol) were suspended in 100 mL dry DMF and de-oxygenated for 20 min. After addition of PdCl2(PPh3)2 (497.7 mg, 0.71 mmol) the mixture was heated to 90 °C for 2 d. The DMF was evaporated and the residue was taken up in 50 mL DCM and washed with H2O repeatedly. After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With pure pentane, then pentane/DCM 3:1 as eluents the pure product was obtained as a colorless crystalline solid (3.50 g, 7.02 mmol, 99%). 1 H NMR (400 MHz, CDCl3): 7.52 (s, 2H, hex), 2.81 (dd, J = 9.2 Hz, 6.8 Hz, 4H, CH2),

166

VI Experimental Part 1.54–1.47 (m, 4H, CH2), 1.36 (d, J = 10.7 Hz, 4H, CH2), 1.34 (s, 24H, CH3), 1.30 (dd, J = 6.3 Hz, 3.5 Hz, 8H, CH2), 0.92–0.86 (m, 6H, CH3).

B–hex–TMS–hex–B TMS

OB O

hex

hex

hex

BO hex O

Br–TMS–Br (170.0 mg, 0.55 mmol), B–hex–B (1.10 g, 2.21 mmol), and Na2CO3 (604.1 mg, 5.70 mmol) were suspended in 30 mL toluene, 10 mL EtOH and 10 mL H2O under N2 atmosphere and the mixture was de-oxygenated for 20 min. After addition of Pd(PPh3)4 (109.8 mg, 0.09 mmol) the reaction mixture was heated to reflux for 3.5 d, cooled to RT, quenched with 100 mL sat. NH4Cl solution and extracted with DCM (3 x 50 mL). The organic phase was dried over Na2SO4, evaporated to dryness and the residue was subjected to gravitational column chromatography. Eluting the product with pentane/DCM 10:1 then 5:1 gave a colorless oil (0.18 g, 0.20 mmol, 37%). 1 H NMR (400 MHz, CDCl3): 7.71 (s, 2H, hex), 7.43 (d, J = 1.7 Hz, 2H, ph), 7.25 (t, J = 1.7 Hz, 1H, ph), 7.09 (s, 2H, hex), 2.92–2.84 (m, 4H, CH2), 2.63–2.56 (m, 4H, CH2), 1.62–1.54 (m, 4H, CH2), 1.49 (dq, J = 12.3 Hz, 7.5 Hz, 4H, CH2), 1.37 (s, 24H, CH3), 1.36–1.29 (m, 12H, CH2), 1.24–1.15 (m, 12H, CH2), 0.91–0.87 (m, 6H, CH3), 0.85–0.80 (m, 6H, CH3), 0.30 (s, 9H, Si(CH3)3).

AQ–bpy–TMS–bpy–AQ TMS

N O

O

N

hex

hex

hex

hex

N N

O

O

B–hex–TMS–hex–B (500.0 mg, 0.56 mmol), AQ–xy–bpy–Br (612.0 mg, 1.12 mmol), and Na2CO3 (356.2 mg, 3.36 mmol) were suspended in 20 mL H2O, 10 mL EtOH, and 50 mL toluene under N2 atmosphere. The suspension was de-oxygenated for 20 min. Following the addition of Pd(PPh3)4 (69.3 mg, 0.06 mmol) the yellow reaction mixture

167

VI Experimental Part was de-oxygenated again for 15 min and heated to reflux for 1.5 d. After addition of 100 mL H2O to the cooled solution, the product was extracted with DCM (3 x 50 mL), dried over Na2SO4 and the solvents were evaporated. Column chromatography on SiO2 with an eluent mixture of DCM + 1% MeOH, then DCM + 3% MeOH gave a yellow solid which was recrystallized from MeOH/hexane (v/v 1:1) (620.0 mg, 0.39 mmol, 71%). 1 H NMR (400 MHz, CDCl3): 8.78 (s, 4H, bpy), 8.61–8.52 (m, 4H, bpy), 8.41 (d, J = 8.0 Hz, 2H, AQ), 8.38–8.35 (m, 6H, AQ), 7.91 (d, J = 8.1 Hz, 4H, bpy), 7.86–7.83 (m, 6H, AQ), 7.54 (d, J = 1.7 Hz, 2H, ph), 7.39 (t, J = 1.7 Hz, 1H, ph), 7.29 (s, 4H, xy/hex), 7.28 (s, 2H, xy/hex), 7.23 (s, 2H, xy/hex), 2.71–2.62 (m, 8H, CH2), 2.38 (d, J = 7.3 Hz, 12H, CH3), 1.58–1.52 (m, 8H, CH2), 1.27–1.19 (m, 24H, CH2), 0.82 (d, J = 5.8 Hz, 12H, CH3), 0.35 (s, 9H, Si(CH3)3).

AQ–bpy–I–bpy–AQ

I

N O

O

N

hex

hex

hex

hex

N N

O

O

AQ–bpy–TMS–bpy–AQ (620.0 mg, 0.39 mmol) was dissolved in 8 mL DCM under N2 atmosphere and cooled to 0 °C. At this temperature a solution of ICl (40.0 µL, 0.79 mmol) in 20 mL MeCN was slowly added dropwise. After stirring at 0 °C for 20 min, the ice-bath was removed and the mixture was stirred at RT over night. The reaction was quenched with sat. aqueous Na2S2O3-solution (100 mL), extracted with DCM (3 x 50 mL), dried over Na2SO4 and the solvents were evaporated. Column chromatography on SiO2 with an eluent mixture of DCM + 2% MeOH gave a yellow solid which was recrystallized from MeOH/hexane (v/v 1:1) (270 mg, 0.17 mmol, 43%). 1 H NMR (400 MHz, CDCl3): 8.77 (dd, J = 2.3 Hz, 0.9 Hz, 2H, bpy), 8.75 (dd, J = 2.3 Hz, 0.9 Hz, 2H, bpy), 8.59–8.52 (m, 4H, bpy), 8.41 (d, J = 8.0 Hz, 2H, AQ), 8.38–8.34 (m, 6H, AQ), 7.89 (ddd, J = 11.1 Hz, 8.1 Hz, 2.3 Hz, 4H, bpy), 7.86–7.82 (m, 6H, AQ), 7.76 (d, J = 1.5 Hz, 2H, ph), 7.37 (t, J = 1.5 Hz, 1H, ph), 7.29 (s, 4H, xy/hex), 7.21 (d, J = 3.5 Hz, 4H, xy/hex), 2.65 (dt, J = 13.2 Hz, 7.9 Hz, 8H, CH2), 2.38 (d, J = 7.2 Hz, 12H, CH3), 1.55–1.48 (m, 8H, CH2), 1.28–1.18 (m, 24H, CH2), 0.87–0.78 (m, 12H, CH3).

168

VI Experimental Part

AQ–bpy–OTA–bpy–AQ

OCH3

OCH3

N

N

H3CO

N O

O

N

OCH3

N hex

hex

hex

hex

N N

O

O

AQ–bpy–I–bpy–AQ (250.0 mg, 0.15 mmol), OTA (115.4 mg, 0.18 mmol), NaOtBu (288.3 mg, 3.00 mmol), Pd(dba)2 (11.5 mg, 0.02 mmol), and (HPtBu3)BF4 (5.80 mg, 0.02 mmol) were suspended in dry, de-oxygenated toluene under N2 atmosphere and heated to reflux for 5 d. After cooling to RT, 100 mL H2O was added and the product was extracted with DCM (3 x 50 mL). The organic phase was dried over Na2SO4, the solvents were evaporated, and the residue was subjected to column chromatography on SiO2. Eluting the product fraction with DCM + 2% MeOH and subsequent recrystallization from hexane gave pure product as a brown solid (200.0 mg, 0.09 mmol, 56%). 1 H NMR (400 MHz, C6D6): 9.07–8.99 (m, 6H, bpy), 8.92 (dd, J = 2.2 Hz, 0.8 Hz, 2H, bpy), 8.50 (d, J = 1.8 Hz, 2H, bpy), 8.36 (d, J = 7.9 Hz, 2H, AQ), 8.32–8.26 (m, 4H, AQ), 7.70 (dd, J = 8.3 Hz, 2.1 Hz, 2H, bpy), 7.60 (d, J = 1.4 Hz, 2H, ph), 7.54–7.50 (m, 4H, AQ), 7.37 (dd, J = 9.2 Hz, 2.1 Hz, 4H, AQ), 7.34 (d, J = 1.7 Hz, 2H, hex/xy), 7.23 (s, 2H, hex/xy), 7.15–7.06 (m, 18H, AA’BB’), 6.99 (d, J = 5.2 Hz, 4H, hex/xy), 6.96 (s, 1H, ph), 6.73–6.68 (m, 6H, AA’BB’), 3.31 (s, 12H, OCH3), 2.92–2.82 (m, 4H, CH2), 2.66–2.57 (m, 4H, CH2), 2.09 (s, 12H, CH3), 1.68–1.61 (m, 4H, CH2), 1.50–1.44 (m, 4H, CH2), 1.27–1.22 (m, 10H, CH2), 1.16–1.05 (m, 14H, CH2), 0.90–0.86 (m, 6H, CH3), 0.80 (t, J = 7.0 Hz, 6H, CH3). ESI-MS calculated (m/z) for C146H138N7O8: 705.6863; found: 705.6866. Singly protonated and triply oxidized (m+H+/3).

169

VI Experimental Part

AQ–Ru–OTA–Ru–AQ

OCH3

OCH3

N

N

H3CO

4+ 4PF6-

OCH3

N hex

hex hex hex

N

N O

N

N Ru

N

Ru N

N O

N

N N

O

N N O

AQ–bpy–OTA–bpy–AQ (130 mg, 0.06 mmol), Ru–pre (87.2 mg, 0.18 mmol), and AgNO3 (183.5 mg, 1.08 mmol) were suspended in 5 mL CHCl3 and 10 mL EtOH under N2 atmosphere. The reaction mixture was de-oxygenated and heated to reflux for 3.5 d. After cooling to RT, 5 mL brine was added, the precipitated AgCl was filtered, washed with DCM, and the solvents were evaporated. The residue was subjected to column chromatography on SiO2. The pentad was purified by two consecutive columns with first acetone/H2O 9:1 + 0.5% sat. aqueous KNO3, then acetone/H2O 20:1 + 1% sat. aqueous KNO3. Sat. aqueous KPF6 solution was added dropwise to an acetone/H2O solution of the pentad to precipitate it as the PF6-salt. This gave the product as a dark red solid. (68.0 mg, 0.02 mmol, 32%). 1 H NMR (400 MHz, CD3CN): 8.63 (dd, J = 8.4 Hz, 6.5 Hz, 4H), 8.55–8.50 (m, 6H), 8.48 (d, J = 8.2 Hz, 2H), 8.34 (d, J = 8.1 Hz, 2H), 8.30 (ddd, J = 9.1 Hz, 4.1 Hz, 2.5 Hz, 4H), 8.21 (d, J = 1.8 Hz, 2H), 8.18 (dd, J = 8.4 Hz, 1.9 Hz, 2H), 8.14–8.02 (m, 9H), 7.99–7.90 (m, 10H), 7.84 (dd, J = 8.0 Hz, 1.8 Hz, 2H), 7.77 (d, J = 5.6 Hz, 4H), 7.68 (d, J = 1.7 Hz, 2H), 7.63 (d, J = 1.7 Hz, 1H), 7.52–6.79 (m, 43H), 3.76 (s, 12H, OCH3), 2.58–2.51 (m, 2H, CH2), 2.28 (s, 6H, CH3), 2.03 (s, 6H, CH3), 1.36–1.28 (m, 4H, CH2), 1.23–1.06 (m, 24H, CH2), 1.02–0.89 (m, 10H, CH2), 0.76 (td, J = 7.0 Hz, 3.5 Hz, 12H, CH3). ESI-MS calculated (m/z) for C186H169N15O8Ru2: 736.0352; found: 736.0361. Elemental analysis calculated (%) for C186H169F24N15O8P4Ru2 · 2 H2O: C 62.75, H 4.90, N 5.90; found: C 62.72, H 5.17, N 5.92.

Sc(HMPA)3 Attention! HMPA is very toxic and presumably a human carcinogen. Sc(OTf)3 was dissolved in dry THF under N2 atmosphere and 3 eq HMPA were added

170

VI Experimental Part dropwise. The colorless, clear solution was heated to 50 °C over night. The white precipitate was filtered, washed with THF, dried, and used without further characterization.

VI.6 Syntheses of the D–bridge–A molecules Br–ph–B

Br

B

Synthesis according to literature procedure.170 1,4-dibromobenzene (1.75 g, 7.42 mmol) was dissolved in 20 mL dry Et2O under N2 atmosphere and cooled to −78 °C. After dropwise addition of n-BuLi (3.40 mL, 5.44 mmol) the mixture was stirred at this temperature for 3 h. A solution of dimesitylboron fluoride (1.56 g, 5.82 mmol) in 5 mL dry Et2O was added dropwise and the suspension was stirred at RT over night. 50 mL Et2O was added and the organic phase was washed with 100 mL sat. NH4Cl-solution and H2O, respectively. After drying over Na2SO4, evaporation of the solvent and column chromatography with pure pentane as eluent, pure product was obtained as a white solid (2.49 g, 6.15 mmol, 80% (of applied 1,4-dibromobenzene)). 1 H NMR (400 MHz, CDCl3): 7.48–7.50 (m, 2H, AA’BB’), 7.36–7.38 (m, 2H, AA’BB’), 6.82 (s, 4H, AA’BB’), 2.31 (s, 6H, CH3), 1.99 (s, 12H, CH3).

Br2–ph–B

Br B Br

1,3,5-tribromobenzene (2.00 g, 6.35 mmol) was dissolved in 30 mL dry Et2O under N2 atmosphere and cooled to −78 °C. After dropwise addition of n-BuLi (2.54 mL, 6.35 mmol) the mixture was stirred at this temperature for 2 h. Dimesitylboron fluoride (1.70 g, 6.35 mmol) was added and the suspension was stirred at RT fro 1 h.

171

VI Experimental Part The suspension was diluted with hexane (100 mL), washed with H2O and dried over Na2SO4. Column chromatography with pure pentane gave pure product as a white solid (2.12 g, 4.38 mmol, 69%). 1 H NMR (400 MHz, CDCl3): 7.76 (t, J = 1.9 Hz, 1H, ph), 7.51 (d, J = 1.9 Hz, 2H, ph), 6.82 (s, 4H, mesityl), 2.31 (s, 6H, CH3), 1.97 (s, 12H, CH3).

Br–ph–B2

B Br B

Br2–ph–B (2.10 g, 4.34 mmol) was treated with n-BuLi (1.74 mL, 4.34 mmol) and dimesitylboron fluoride (1.16 g, 4.34 mmol) in the same way as described for the preparation of Br2–ph–B(mes)2. After purification by column chromatography with pentane as eluent product was obtained as a white solid (2.12 g, 4.38 mmol, 69%). 1 H NMR (400 MHz, CDCl3): 7.65 (d, J = 1.1 Hz, 2H, ph), 7.38 (t, J = 1.1 Hz, 1H, ph), 6.75 (s, 8H, mesityl), 2.28 (s, 12H, CH3), 1.94 (s, 24H, CH3).

Br–ph–acetylene–TMS

Br

TMS

Synthesis by adaption of literature procedure.171 1-bromo-4-iodobenzene (2.00 g, 7.06 mmol) and trimethylsilyl acetylene (1.10 mL, 7.78 mmol) were dissolved in 30 mL dry NEt3 under N2 atmosphere and the mixture was thoroughly de-oxygenated. PdCl2(PPh3)2 (98.2 mg, 0.14 mmol) and CuI (53.2 mg, 0.28 mmol) were added and the reaction mixture was heated to 100 °C for 30 min. After cooling to RT 100 mL sat. NH4Cl-solution was added, the product was extracted with DCM (3 x 50 mL), the organic phases were dried over Na2SO4 and the solvent was evaporated. Column chromatography with pure pentane as eluent gave the pure product as a white solid (1.80 g, 7.11 mmol, quant.). 1 H NMR (400 MHz, CDCl3): 7.41–7.45 (m, 2H, AA’BB’), 7.30–7.33 (m, 2H, AA’BB’), 0.25 (s, 9H, Si(CH3)3).

172

VI Experimental Part

OTA–ph–acetylene–TMS OCH3

H3CO

N

TMS

N

H3CO

N

OCH3

Br–ph–acetylene–TMS (150 mg, 0.59 mmol), OTA (307.9 mg, 0.49 mmol), NaOtBu (941.8 mg, 9.80 mmol), Pd(dba)2 (28.2 mg, 0.05 mmol), and (HPtBu3)BF4 (14.2 mg, 0.05 mmol) were dissolved in 15 mL dry, de-oxygenated toluene under N2 atmosphere. The mixture was stirred at 120 °C for 20 h. 100 mL brine was added to the cool mixture and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With DCM as eluent the pure product was obtained as a yellow oil which solidifies with time (290 mg, 0.36 mmol, 74%). 1 H NMR (400 MHz, acetone d6): 7.23–7.25 (m, 2H, AA’BB’), 7.0–7.06 (m, 8H, AA’BB’), 6.97–6.99 (m, 4H, AA’BB’), 6.82–6.90 (m, 14H, AA’BB’), 3.78 (s, 12H, OCH3), 0.19 (s, 9H, Si(CH3)3).

OTA–ph–acetylene–H OCH3

H3CO

N

N

H3CO

N

OCH3

OTA–ph–acetylene–TMS (300.0 mg, 0.38 mmol) was dissolved in 10 mL dry THF under N2 atmosphere and a TBAF-solution in THF (1.40 mL, 0.38 mmol) was added dropwise. The mixture was stirred at RT for 1.5 h. After removal of the THF, the residue was taken up in EtOAc (100 mL) and washed with H2O repeatedly. The organic phases were dried over Na2SO4 and evaporated to dryness which gave pure

173

VI Experimental Part product as a yellow solid (250.0 mg, 0.35 mmol, 91%). 1 H NMR (400 MHz, C6D6): 7.38–7.36 (m, 2H, AA’BB’), 7.12–7.07 (m, 8H, AA’BB’), 7.05–6.97 (m, 10H, AA’BB’), 6.73–6.68 (m, 8H, AA’BB’), 3.29 (s, 12H, OCH3), 2.76 (s, 1H, acetylene).

OTA–long–B

OCH3

H3CO

N

N

H3CO

B

N

OCH3

OTA–ph–acetylene–H (217.2 mg, 0.30 mmol) and Br–ph–B (100 mg, 0.25 mmol) were dissolved in 10 mL dry NEt3 under N2 atmosphere and the mixture was deoxygenated for 20 min. PdCl2(PPh3)2 (3.51 mg, 0.005 mmol) and CuI (1.90 mg, 0.01 mmol) were added and the reaction mixture was heated to 100 °C over night. After cooling to RT 100 mL sat. NH4Cl-solution was added, the product was extracted with DCM (3 x 50 mL), the organic phases were dried over Na2SO4 and the solvent was evaporated. Column chromatography with an eluent mixture of pentane/DCM 1:2 yielded a bright orange oil which was crystallized as an orange solid by dropping a concentrated DCM-solution into H2O (111.0 mg, 0.11 mmol, 42%). 1 H NMR (400 MHz, acetone d6): 7.50–7.52 (d, J = 8.0 Hz, 2H, AA’BB’), 7.44–7.46 (d, J = 7.9 Hz, 2H, AA’BB’), 7.33–7.37 (m, 2H, AA’BB’), 7.00–7.07 (m, 12H, AA’BB’), 6.84– 6.91 (m, 18H, AA’BB’ + mesityl), 3.77 (s, 12H, OCH3), 2.28 (s, 6H, CH3), 2.00 (s, 12H, CH3). 13C NMR (100 MHz, d6 acetone): 157.2, 150.2, 146.7, 146.4, 142.5, 141.9, 141.5, 140.4, 139.8, 139.1, 137.1, 133.6, 131.7, 129.3, 127.9, 127.5, 122.4, 119.8, 115.7, 114.1, 93.8, 89.3, 55.9, 23.9, 21.4. ESI-MS calculated (m/z) for C72H66BN3O4: 1047.5152; found: 1047.5127. Elemental analysis calculated (%) for C72H66BN3O4: C 82.51, H 6.35, N 4.01; found: C 82.45, H 6.65, N 3.80.

174

VI Experimental Part

OTA–long–B2

OCH3

H3CO

N B N B

H3CO

N

OCH3

OTA–ph–acetylene–H (180.0 mg, 0.25 mmol) and Br–ph–B2 (194.9 mg, 0.29 mmol) were suspended in 10 mL dry NEt3 under N2 atmosphere and the mixture was de-oxygenated for 10 min. PdCl2(PPh3)2 (3.50 mg, 0.005 mmol) and CuI (1.90 mg, 0.01 mmol) were added and the reaction mixture was heated to 100 °C over night. After cooling to RT 100 mL sat. NH4Cl-solution was added, the product was extracted with DCM (3 x 50 mL), the organic phases were dried over Na2SO4 and the solvent was evaporated. Column chromatography with an eluent mixture of pentane/DCM 1:2 yielded a bright yellow oil, which was crystallized as an orange solid by dropping a concentrated DCM-solution into MeOH (110.0 mg, 0.09 mmol, 34%). 1 H NMR (400 MHz, acetone d6): 7.66 (d, J = 1.3 Hz, 2H, ph), 7.54 (t, J = 1.3 Hz, 1H, ph), 7.25 (d, J = 8.7 Hz, 2H, AA’BB’), 7.07–7.01 (m, 8H, AA’BB’), 6.98 (d, J = 8.7 Hz, 4H, AA’BB’), 6.92–6.86 (m, 8H, AA’BB’), 6.86–6.81 (m, 6H, AA’BB’), 6.79 (s, 8H, mesityl), 3.77 (s, 12H, OCH3), 2.24 (s, 12H, CH3), 1.97 (s, 24H, CH3). 13C NMR (100 MHz, d6 acetone): 157.2, 149.9, 148.0, 146.7, 142.5, 142.2, 141.9, 141.5, 140.5, 140.2, 133.4, 129.3, 127.7, 127.4, 124.8, 122.4, 119.8, 115.8, 115.2, 114.2, 91.6, 88.9, 55.9, 23.9, 21.5. ESI-MS calculated (m/z) for C90H87B2N3O4: 1295.6902; found: 1295.6878. Elemental analysis calculated (%) for C90H87B2N3O4 · 0.5 H2O: C 82.81, H 6.80, N 3.22; found: C 82.64, H 6.74, N 2.94.

175

VI Experimental Part

OTA–short–B OCH3

H3CO

N

B

N

H3CO

N

OCH3

para-Br–ph–B (71.5 mg, 0.18 mmol) and OTA (100.0 mg, 0.16 mmol), together with NaOtBu (307.6 mg, 3.20 mmol), Pd(dba)2 (11.5 mg, 0.02 mmol), and (HPtBu3)BF4 (4.60 mg, 0.02 mmol) were dissolved in 10 mL dry, de-oxygenated toluene under N2 atmosphere. The mixture was refluxed at 125 °C for 1 h and 15 min and cooled to RT. 100 mL brine was added and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With pentane/DCM 1:1 then pure DCM as eluents the orange luminescent fraction was collected and recrystallized from acetone to yield a bright yellow solid (117 mg, 0.12 mmol, 77%). 1 H NMR (400 MHz, acetone d6): 7.31–7.25 (m, 2H, AA’BB’), 7.11–7.02 (m, 12H, AA’BB’), 6.93–6.84 (m, 12H, AA’BB’), 6.82–6.76 (m, 6H, AA’BB’), 3.78 (s, 12H, OCH3), 2.25 (s, 6H, CH3), 2.04 (s, 12H, CH3; signal under acetone). 13C NMR (100 MHz, C6D6): 156.2, 152.5, 146.1, 142.4, 141.2, 140.7, 139.5, 137.6, 136.2, 128.4, 127.6, 126.5, 125.7, 121.9, 117.1, 114.9, 54.7, 23.6, 21.0. ESI-MS calculated (m/z) for C64H62N3BO4: 947.4838; found: 947.4834. Elemental analysis calculated (%) for C64H62N3BO4 · 0.5 H2O: C 80.32, H 6.64, N 4.39; found: C 80.36, H 6.81, N 4.46.

OTA–short–B2 OCH3

H3CO

N B N B

H3CO

N

OCH3

176

VI Experimental Part Br–ph–B2 (117.6 mg, 0.18 mmol), OTA (100.0 mg, 0.16 mmol), NaOtBu (307.6 mg, 3.20 mmol), Pd(dba)2 (11.5 mg, 0.02 mmol), and (HPtBu3)BF4 (4.60 mg, 0.02 mmol) were dissolved in 10 mL dry, de-oxygenated toluene under N2 atmosphere. The mixture was refluxed at 125 °C for 1 h and cooled to RT. 100 mL brine was added and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With pentane/DCM 1:1 then pure DCM as eluents the orange fraction was collected and concentrated. Pure, orange product was obtained by dropping a concentrated DCM-solution in MeOH (120.0 mg, 0.10 mmol, 63%). 1 H NMR (400 MHz, acetone d6): 7.27 (d, J = 1.1 Hz, 2H, ph), 7.22 (t, J = 1.1 Hz, 1H, ph), 6.97–6.92 (m, 8H, AA’BB’), 6.90–6.84 (m, 12H, AA’BB’), 6.79–6.74 (m, 12H, AA’BB’ + mesityl), 3.79 (s, 12H, OCH3), 2.22 (s, 12H, CH3), 1.98 (s, 24H, CH3). 13C NMR (100 MHz, d6 acetone): 156.8, 145.1, 142.8, 142.3, 141.4, 139.7, 138.4, 136.8, 134.2, 129.2, 127.8, 126.7, 125.7, 123.6, 120.1, 115.6, 55.9, 23.8, 21.6. ESI-MS calculated (m/z) for C82H83B2N3O4: 1195.6588; found: 1195.6576. Elemental analysis calculated (%) for C82H83B2N3O4 · 0.5 H2O: C 81.72, H 7.02, N 3.49; found: C 81.83, H 6.95, N 3.44.

TAA–ph–acetylene–TMS

H3CO

N

TMS

H3CO

Br–ph–acetylene–TMS (1.00 g, 3.95 mmol), dianisylamine (724.5 mg, 3.16 mmol), NaOtBu (6.30 g, 65.56 mmol), Pd(dba)2 (190.0 mg, 0.33 mmol), and (PtBu3 (0.98 mL, 0.33 mmol) were dissolved in 60 mL dry, de-oxygenated toluene under N2 atmosphere. The mixture was stirred at 120 °C for 30 h. 100 mL brine was added to the cool mixture and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With pentane/DCM 1:3 as eluent mixture the pure product was obtained as a yellow solid (810.0 mg, 2.02 mmol, 64%). 1 H NMR (250 MHz, acetone d6): 7.27–7.19 (m, 2H, AA’BB’), 7.12–7.05 (m, 4H, AA’BB’), 6.97–6.89 (m, 4H, AA’BB’), 6.76–6.68 (m, 2H, AA’BB’), 3.80 (s, 6H, OCH3), 0.20 (s, 9H, Si(CH3)3).

177

VI Experimental Part

TAA–ph–acetylene–H H3CO

N

H3CO

TAA–ph–acetylene–TMS (0.80 g, 1.99 mmol) was dissolved in 40 mL THF under N2 atmosphere and a TBAF-solution in THF (18.0 mL, 5.40 mmol) was added dropwise. After the mixture was stirred at RT for 1 h, THF was evaporated and the residue was taken up in EtOAc and washed with H2O (3 x 100 mL). After drying over Na2SO4 and evaporation of the solvent pure product was obtained as a yellow solid (0.65 g, 1.97 mmol, 99%). 1 H NMR (250 MHz, CDCl3): 7.30–7.23 (m, 2H, AA’BB’), 7.11–7.01 (m, 4H, AA’BB’), 6.89–6.76 (m, 6H, AA’BB’), 3.80 (s, 6H, OCH3).

TAA–long–B H3CO

N

B

H3CO

A suspension of para-Br–ph–B (270.0 mg, 0.67 mmol) and TAA–ph–acetylene–H (200.0 mg, 0.61 mmol) in 15 mL dry NEt3 under N2 atmosphere was de-oxygenated for 30 min. CuI (4.50 mg, 0.02 mmol) and PdCl2(PPh3)2 (8.40 mg, 0.01 mmol) were added and the reaction mixture was stirred at 100 °C for 30 h. After cooling to RT 100 mL sat. NH4Cl solution was added and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With pentane/EtOAc 18:1 the product fraction was eluted and then concentrated. Subsequent crystallization by dropping a concentrated DCM-solution into MeOH yielded pure product as a yellow solid (172.0 mg, 0.26 mmol, 43%). 1 H NMR (250 MHz, acetone d6): 7.56–7.49 (m, 2H, AA’BB’), 7.49–7.42 (m, 2H, AA’BB’), 7.39–7.32 (m, 2H, AA’BB’), 7.17–7.08 (m, 4H, AA’BB’), 6.99–6.91 (m, 4H, AA’BB’), 6.85 (s, 4H, mesityl), 6.82–6.75 (m, 2H, AA’BB’), 3.81 (s, 6H, OCH3), 2.29 (s, 6H, CH3), 2.00 (s, 12H, CH3). 13C NMR (100 MHz, d6 acetone): 158.9, 151.5, 147.3,

178

VI Experimental Part 143.4, 142.5, 141.6, 140.8, 138.0, 134.5, 132.7, 130.2, 129.5, 129.4, 120.0, 116.9, 114.8, 94.7, 90.1, 56.9, 24.8, 22.4. ESI-MS calculated (m/z) for C46H44BNO2: 653.3467; found: 653.3456. Elemental analysis calculated (%) for C46H44BNO2 · 0.5 H2O: C 83.37, H 6.84, N 2.11; found: C 83.74, H 6.83, N 2.20.

TAA–short–B H3CO

N

B

H3CO

para-Br–ph–B (0.20 g, 0.51 mmol), dianisylamine (0.11 g, 0.47 mmol), NaOtBu (0.82 g, 8.53 mmol), Pd(dba)2 (23.0 mg, 0.04 mmol), and (HPtBu3)BF4 (8.30 mg, 0.04 mmol) were dissolved in 15 mL dry, de-oxygenated toluene under N2 atmosphere. The mixture was stirred at 120 °C for 2 h. 100 mL brine was added to the cool mixture and the aqueous phase was extracted with DCM (3 x 50 mL). After drying over Na2SO4 and evaporation of the solvents the crude product was subjected to column chromatography. With pentane/EtOAc 18:1 as eluent mixture the pure product was obtained as a bright yellow solid (0.24 g, 0.43 mmol, 93%). 1 H NMR (250 MHz, DMSO d6): 7.21–7.10 (m, 6H, AA’BB’), 6.99–6.91 (m, 4H, AA’BB’), 6.77 (s, 4H, mesityl), 6.61 (d, J = 8.7 Hz, 2H, AA’BB’), 3.74 (s, 6H, OCH3), 2.22 (s, 6H, CH3), 1.97 (s, 12H, CH3). 13C NMR (100 MHz, DMSO d6): 156.7, 152.1, 141.4, 139.7, 138.5, 138.2, 137.2, 134.3, 128.1, 127.9, 115.3, 115.0, 55.2, 23.0, 20.7. ESI-MS calculated (m/z) for C38H40BNO2: 553.3153; found: 553.3147. Elemental analysis calculated (%) for C38H40BNO2: C 82.45, H 7.28, N 2.53; found: C 82.05, H 7.29, N 2.40.

179

VII. Bibliography [1] Department of Economic and Social Affairs, Population Division World Population Prospects: The 2012 Revision, Highlights and Advance Tables. Tech. rep.; United Nations, 2013. [2] Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci. U S A. 2006, 103, 15729. [3] Crabtree, G. W.; Lewis, N. S. Physics Today 2007, 60, 37. [4] Ciamician, G. Science 1912, 36, 385. [5] Balzani, V.; Moggi, L.; Manfrin, M. F.; Bolletta, F.; Gleria, M. Science 1975, 189, 852. [6] Swierk, J. R.; Mallouk, T. E. Chem. Soc. Rev. 2013, 42, 2357. [7] Foyer, C. H.; Bloom, A. J.; Queval, G.; Noctor, G. Annu. Rev. Plant Biol. 2009, 60, 455. [8] Das, D.; Vezirolu, T. Int. J. Hydrogen Energ. 2001, 26, 13. [9] Zhu, M.; Yang, C. Chem. Soc. Rev. 2013, 42, 4963. [10] Reineke, S.; Thomschke, M.; Lüssem, B.; Leo, K. Rev. Mod. Phys. 2013, 85, 1245. [11] Tachibana, Y.; Vayssieres, L.; Durrant, J. R. Nat. Photon. 2012, 6, 511. [12] Barber, J. Chem. Soc. Rev. 2009, 38, 185. [13] Kärkäs, M. D.; Johnston, E. V.; Verho, O.; Åkermark, B. Acc. Chem. Res. 2014, 47, 100. [14] Pace, R. J.; Stranger, R.; Petrie, S. Dalton Trans. 2012, 41, 7179. [15] Umena, Y.; Kawakami, K.; Shen, J.-R.; Kamiya, N. Nature 2011, 473, 55. [16] Barber, J. Inorg. Chem. 2008, 47, 1700. [17] Loll, B.; Kern, J.; Saenger, W.; Zouni, A.; Biesiadka, J. Nature 2005, 438, 1040. [18] Dau, H.; Haumann, M. Coord. Chem. Rev. 2008, 252, 273. [19] Haumann, M.; Liebisch, P.; Müller, C.; Barra, M.; Grabolle, M.; Dau, H. Science 2005, 310, 1019.

181

VII Bibliography [20] Esswein, A. J.; Nocera, D. G. Chem. Rev. 2007, 107, 4022. [21] Elvington, M.; Brown, J.; Arachchige, S. M.; Brewer, K. J. J. Am. Chem. Soc. 2007, 129, 10644. [22] Eckenhoff, W. T.; Eisenberg, R. Dalton Trans. 2012, 41, 13004. [23] Dau, H.; Limberg, C.; Reier, T.; Risch, M.; Roggan, S.; Strasser, P. ChemCatChem 2010, 2, 724. [24] Fujishima, A.; Honda, K. Nature 1972, 238, 37. [25] Gersten, S. W.; Samuels, G. J.; Meyer, T. J. J. Am. Chem. Soc. 1982, 104, 4029. [26] Gao, Y.; Duan, L.; Yu, Z.; Ding, X.; Sun, L. Faraday Discuss. 2014, 176, 225. [27] Brimblecombe, R.; Koo, A.; Dismukes, G. C.; Swiegers, G. F.; Spiccia, L. J. Am. Chem. Soc. 2010, 132, 2892. [28] Cape, J. L.; Hurst, J. K. J. Am. Chem. Soc. 2008, 130, 827. [29] Du, P.; Eisenberg, R. Energy Environ. Sci. 2012, 5, 6012. [30] Morris, A. J.; Meyer, G. J.; Fujita, E. Acc. Chem. Res. 2009, 42, 1983. [31] Tinnemans, A. H. A.; Koster, T. P. M.; Thewissen, D. H. M. W.; Mackor, A. Recl. Trav. Chim. Pays-Bas 1984, 103, 288. [32] Sato, S.; Koike, K.; Inoue, H.; Ishitani, O. Photochem. Photobiol. Sci. 2007, 6, 454. [33] Atkins, P. W.; de Paula, J., Physikalische Chemie, Fourth; Wiley-VCH: Weinheim, 2006. [34] Lakowicz, J. R., Principles of Fluorescence Spectroscopy; Kluwer Academic/Plenum Publishers: New York, 1999; Vol. 2. [35] Montalti, M.; Credi, A.; Prodi, L.; Gandolfi, M. T., Handbook of Photochemistry; CRC Press: Florida, 2006; Vol. 3. [36] Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85. [37] Karlsson, S. Single and Accumulative Electron Transfer - Prerequisites for Artificial Photosynthesis., Ph.D. Thesis, Uppsala Universitet, 2010. [38] Marcus, R. A. J. Chem. Phys. 1956, 24, 966. [39] Marcus, R. A. J. Chem. Phys. 1957, 26, 867. [40] Marcus, R. A. J. Chem. Phys. 1957, 26, 872.

182

VII Bibliography [41] Marcus, R. A. Discuss. Faraday Soc. 1960, 29, 21. [42] Marcus, R. A. J. Phys. Chem. 1963, 67, 853. [43] Marcus, R. A. Annu. Rev. Phys. Chem. 1964, 15, 155. [44] Marcus, R.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265. [45] Hush, N. S. Trans. Faraday Soc. 1961, 57, 557. [46] Miller, J. R.; Calcaterra, L. T.; Closs, G. L. J. Am. Chem. Soc. 1984, 106, 3047. [47] Closs, G. L.; Miller, J. R. Science 1988, 240, 440. [48] Creutz, C.; Taube, H. J. Am. Chem. Soc. 1969, 91, 3988. [49] Edwards, P.; Gray, H.; Lodge, M.; Williams, R. Angew. Chem. 2008, 120, 6860. [50] Förster, T. Discuss. Faraday Soc. 1959, 27, 7. [51] Dexter, D. L. J. Chem. Phys. 1953, 21, 836. [52] Karlsson, S.; Boixel, J.; Pellegrin, Y.; Blart, E.; Becker, H.-C.; Odobel, F.; Hammarström, L. Faraday Discuss. 2012, 155, 233. [53] Naqvi, K. R.; Staerk, H.; Gillbro, T. Chem. Phys. Lett. 1977, 49, 160. [54] Ceroni, P.; Credi, A.; Venturi, M.; Balzani, V. Photochem. Photobiol. Sci. 2010, 9, 1561. [55] Pellegrin, Y.; Odobel, F. Coord. Chem. Rev. 2011, 255, 2578. [56] Manbeck, G. F.; Brewer, K. J. Coord. Chem. Rev. 2013, 257, Electron Transfer in Coordination Chemistry, 1660. [57] Bonn, A. G.; Wenger, O. S. Chimia 2015, 69, 17. [58] O’Neil, M. P.; Niemczyk, M. P.; Svec, W. A.; Gosztola, D.; Gaines, G. L.; Wasielewski, M. R. Science 1992, 257, 63. [59] Imahori, H.; Hasegawa, M.; Taniguchi, S.; Aoki, M.; Okada, T.; Sakata, Y. Chem. Lett. 1998, 27, 721. [60] Molnar, S. M.; Nallas, G.; Bridgewater, J. S.; Brewer, K. J. J. Am. Chem. Soc. 1994, 116, 5206. [61] Elvington, M.; Brewer, K. J. Inorg. Chem. 2006, 45, 5242. [62] Arachchige, S. M.; Brown, J.; Brewer, K. J. J. Photochem. Photobiol., A 2008, 197, 13. [63] Zigler, D. F.; Wang, J.; Brewer, K. J. Inorg. Chem. 2008, 47, 11342.

183

VII Bibliography [64] Arachchige, S. M.; Brown, J. R.; Chang, E.; Jain, A.; Zigler, D. F.; Rangan, K.; Brewer, K. J. Inorg. Chem. 2009, 48, 1989. [65] White, T. A.; Rangan, K.; Brewer, K. J. J. Photo 2010, 209, 203. [66] Konduri, R.; Ye, H.; MacDonnell, F. M.; Serroni, S.; Campagna, S.; Rajeshwar, K. Angew. Chem. Int. Ed. 2002, 41, 3185. [67] Konduri, R.; de Tacconi, N. R.; Rajeshwar, K.; MacDonnell, F. M. J. Am. Chem. Soc. 2004, 126, 11621. [68] Ghaddar, T. H.; Wishart, J. F.; Thompson, D. W.; Whitesell, J. K.; Fox, M. A. J. Am. Chem. Soc. 2002, 124, 8285. [69] Zhang, X.-F.; Cui, X.; Liu, Q.; Zhang, F. ChemPhysChem 2008, 9, 1514. [70] Zhang, X.-F.; Cui, X.; Liu, Q.; Zhang, F. Phys. Chem. Chem. Phys. 2009, 11, 3566. [71] Elliott, K. J.; Harriman, A.; Le Pleux, L.; Pellegrin, Y.; Blart, E.; Mayer, C. R.; Odobel, F. Phys. Chem. Chem. Phys. 2009, 11, 8767. [72] Matt, B.; Fize, J.; Moussa, J.; Amouri, H.; Pereira, A.; Artero, V.; Izzet, G.; Proust, A. Energy Environ. Sci. 2013, 6, 1504. [73] Sun, L.; Raymond, M. K.; Magnuson, A.; LeGourriérec, D.; Tamm, M.; Abrahamsson, M.; Kenéz, P. H.; Mårtensson, J.; Stenhagen, G.; Hammarström, L.; Styring, S.; Åkermark, B. J. Inorg. Biochem. 2000, 78, 15. [74] Huang, P.; Magnuson, A.; Lomoth, R.; Abrahamsson, M.; Tamm, M.; Sun, L.; van Rotterdam, B.; Park, J.; Hammarström, L.; Åkermark, B.; Styring, S. J. Inorg. Biochem. 2002, 91, 159. [75] Romain, S.; Baffert, C.; Dumas, S.; Chauvin, J.; Lepretre, J.-C.; Daveloose, D.; Deronzier, A.; Collomb, M.-N. Dalton Trans. 2006, 5691. [76] Collomb, M.-N.; Deronzier, A. Eur. J. Inorg. Chem. 2009, 2009, 2025. [77] Karlsson, S.; Boixel, J.; Pellegrin, Y.; Blart, E.; Becker, H.-C.; Odobel, F.; Hammarström, L. J. Am. Chem. Soc. 2010, 132, 17977. [78] Gray, H. B.; Maverick, A. W. Science 1981, 214, 1201. [79] Meyer, T. J. Acc. Chem. Res. 1989, 22, 163. [80] Sacksteder, L.; Zipp, A. P.; Brown, E. A.; Streich, J.; Demas, J. N.; DeGraff, B. A. Inorg. Chem. 1990, 29, 4335. [81] Chen, P.; Westmoreland, T. D.; Danielson, E.; Schanze, K. S.; Anthon, D.; Neveux, P. E.; Meyer, T. J. Inorg. Chem. 1987, 26, 1116.

184

VII Bibliography [82] Kalyanasundaram, K. Coord. Chem. Rev. 1982, 46, 159. [83] Caspar, J. V.; Meyer, T. J. J. Am. Chem. Soc. 1983, 105, 5583. [84] Caspar, J. V.; Meyer, T. J. J. Phys. Chem. 1983, 87, 952. [85] Wenger, O. S. Coord. Chem. Rev. 2009, 253, 1439. [86] Diaz, A. N. J. Photochem. Photobiol., A 1990, 53, 141. [87] Görner, H. Photochem. Photobiol. 2003, 77, 171. [88] Hankache, J.; Hanss, D.; Wenger, O. S. J. Phys. Chem. A 2012, 116, 3347. [89] Hankache, J.; Wenger, O. S. Phys. Chem. Chem. Phys. 2012, 14, 2685. [90] Babaei, A.; Connor, P. A.; McQuillan, A. J.; Umapathy, S. J. Chem. Educ. 1997, 74, 1200. [91] Das, S.; Bhattacharya, A.; Mandal, P.; Rath, M.; Mukherjee, T. Radiat. Phys. Chem. 2002, 65, 93. [92] Rath, M.; Pal, H.; Mukherjee, T. Radiat. Phys. Chem. 1996, 47, 221. [93] Wightman, R. M.; Cockrell, J. R.; Murray, R. W.; Burnett, J. N.; Jones, S. B. J. Am. Chem. Soc. 1976, 98, 2562. [94] Hirao, Y.; Ito, A.; Tanaka, K. J. Phys. Chem. A 2007, 111, 2951. [95] Bozic-Weber, B.; Brauchli, S. Y.; Constable, E. C.; Furer, S. O.; Housecroft, C. E.; Wright, I. A. Phys. Chem. Chem. Phys. 2013, 15, 4500. [96] Ito, A.; Sakamaki, D.; Ichikawa, Y.; Tanaka, K. Chem. Mater. 2011, 23, 841. [97] Hirao, Y.; Ino, H.; Ito, A.; Tanaka, K.; Kato, T. J. Phys. Chem. A 2006, 110, 4866. [98] Kamino, B. A.; Morse, G. E.; Bender, T. P. J. Phys. Chem. C 2011, 115, 20716. [99] Ning, Z.; Tian, H. Chem. Commun. 2009, 5483. [100] Amthor, S.; Noller, B.; Lambert, C. Chem. Phys. 2005, 316, 141. [101] Smieja, J. M.; Kubiak, C. P. Inorg. Chem. 2010, 49, 9283. [102] Sullivan, B. P.; Meyer, T. J. J. Chem. Soc., Chem. Commun. 1984, 1244. [103] Vlˇcek A., J. English In Photophysics of Organometallics, Lees, A. J., Ed.; Topics in Organometallic Chemistry, Vol. 29; Springer Berlin Heidelberg: 2010, p 115. [104] Pavlishchuk, V. V.; Addison, A. W. Inorg. Chim. Acta 2000, 298, 97. [105] Kwon, O.; Barlow, S.; Odom, S. A.; Beverina, L.; Thompson, N. J.; Zojer, E.; Brédas, J.-L.; Marder, S. R. J. Phys. Chem. A 2005, 109, 9346.

185

VII Bibliography [106] Weller, A. Z. Phys. Chem. 1982, 133, 93. [107] Hori, H.; Ishihara, J.; Koike, K.; Takeuchi, K.; Ibusuki, T.; Ishitani, O. J. Photochem. Photobiol., A 1999, 120, 119. [108] Kuss-Petermann, M.; Wolf, H.; Stalke, D.; Wenger, O. S. J. Am. Chem. Soc. 2012, 134, 12844. [109] Chattopadhyay, N.; Serpa, C.; Purkayastha, P.; Arnaut, L. G.; Formosinho, S. J. Phys. Chem. Chem. Phys. 2001, 3, 70. [110] Pawlicki, M.; Collins, H.; Denning, R.; Anderson, H. Angew. Chem. Int. Ed. 2009, 48, 3244. [111] Pollnau, M.; Gamelin, D. R.; Lüthi, S. R.; Güdel, H. U.; Hehlen, M. P. Phys. Rev. B 2000, 61, 3337. [112] Sreenath, K.; Thomas, T. G.; Gopidas, K. R. Org. Lett. 2011, 13, 1134. [113] Hankache, J.; Wenger, O. S. Chem. Commun. 2011, 47, 10145. [114] Kalyanasundaram, K. J. Chem. Soc., Faraday Trans. 2 1986, 82, 2401. [115] Záliš, S.; Consani, C.; Nahhas, A. E.; Cannizzo, A.; Chergui, M.; Hartl, F.; Jr., A. V. Inorg. Chim. Acta 2011, 374, 578. [116] Protti, S.; Fagnoni, M.; Monti, S.; Rehault, J.; Poizat, O.; Albini, A. RSC Adv. 2012, 2, 1897. [117] Pierri, A. E.; Pallaoro, A.; Wu, G.; Ford, P. C. J. Am. Chem. Soc. 2012, 134, 18197. [118] Fang, Y.-Q.; Hanan, G. S. Synlett 2003, 6, 852. [119] Freys, J. C.; Wenger, O. S. Eur. J. Inorg. Chem. 2010, 35, 5509. [120] Liu, S.-J.; Zhao, Q.; Chen, R.-F.; Deng, Y.; Fan, Q.-L.; Li, F.-Y.; Wang, L.-H.; Huang, C.-H.; Huang, W. Chem. Eur. J. 2006, 12, 4351. [121] So, K. H.; Kim, R.; Park, H.; Kang, I.; Thangaraju, K.; Park, Y. S.; Kim, J. J.; Kwon, S.-K.; Kim, Y.-H. Dyes Pigm. 2012, 92, 603. [122] Inoue, H.; Hida, M.; Nakashima, N.; Yoshihara, K. J. Phys. Chem. 1982, 86, 3184. [123] Lin, Z.-P.; Aue, W. A. Spectrochim. Acta, Part A 2000, 56, 111. [124] Katsumi, J.; Nakayama, T.; Esaka, Y.; Uno, B. Anal. Sci. 2012, 28, 257. [125] Raamat, E.; Kaupmees, K.; Ovsjannikov, G.; Trummal, A.; Kütt, A.; Saame, J.; Koppel, I.; Kaljurand, I.; Lipping, L.; Rodima, T.; Pihl, V.; Koppel, I. A.; Leito, I. J. Phys. Org. Chem. 2013, 26, 162.

186

VII Bibliography [126] Kütt, A.; Rodima, T.; Saame, J.; Raamat, E.; Mäemets, V.; Kaljurand, I.; Koppel, I. A.; Garlyauskayte, R. Y.; Yagupolskii, Y. L.; Yagupolskii, L. M.; Bernhardt, E.; Willner, H.; Leito, I. J. Org. Chem. 2011, 76, 391. [127] Howells, R. D.; Mc Cown, J. D. Chem. Rev. 1977, 77, 69. [128] Kantner, S. S.; Kreevoy, M. M. J. Org. Chem. 1977, 42, 865. [129] Kütt, A.; Leito, I.; Kaljurand, I.; Sooväli, L.; Vlasov, V. M.; Yagupolskii, L. M.; Koppel, I. A. J. Org. Chem. 2006, 71, 2829. [130] Sarmini, K.; Kenndler, E. J. Biochem. Biophys. Methods 1999, 38, 123. [131] Gupta, N.; Linschitz, H. J. Am. Chem. Soc. 1997, 119, 6384. [132] Hankache, J.; Niemi, M.; Lemmetyinen, H.; Wenger, O. S. Inorg. Chem. 2012, 51, 6333. [133] Fukuzumi, S.; Ohkubo, K.; Morimoto, Y. Phys. Chem. Chem. Phys. 2012, 14, 8472. [134] Fukuzumi, S.; Ohkubo, K. Coord. Chem. Rev. 2010, 254, Inorganic Reaction Mechanisms A Tribute to Ralph Pearson on the occasion of his 90th birthday, 372. [135] Fukuzumi, S.; Ohtsu, H.; Ohkubo, K.; Itoh, S.; Imahori, H. Coord. Chem. Rev. 2002, 226, 71. [136] Okamoto, K.; Imahori, H.; Fukuzumi, S. J. Am. Chem. Soc. 2003, 125, 7014. [137] Shannon, R. D. Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Cryst., 1976, 32, 751. [138] Kawashima, T.; Ohkubo, K.; Fukuzumi, S. Phys. Chem. Chem. Phys. 2011, 13, 3344. [139] Donoghue, J.; Fernandez, E.; McMillan, J.; Peters, D. J. Inorg. Nucl. Chem. 1969, 31, 1431. [140] Greaves, M. D.; Niemz, A.; Rotello, V. M. J. Am. Chem. Soc. 1999, 121, 266. [141] Hatchard, C. G.; Parker, C. A. Proc. R. Soc. Math. Phys. Eng. Sci. 1956, 235, 518. [142] Pozdnyakov, I. P.; Kel, O. V.; Plyusnin, V. F.; Grivin, V. P.; Bazhin, N. M. J. Phys. Chem. A 2008, 112, 8316. [143] Vitz, E. W. J. Chem. Educ. 1981, 58, 655. [144] Achord, J. M.; Hussey, C. L. Anal. Chem. 1980, 52, 601.

187

VII Bibliography [145] Nissim, R.; Batchelor-McAuley, C.; Li, Q.; Compton, R. G. J. Electroanal. Chem. 2012, 681, 44. [146] Breslin, D. T.; Fox, M. A. J. Org. Chem. 1994, 59, 7557. [147] Reynolds, R.; Line, L. L.; Nelson, R. F. J. Am. Chem. Soc. 1974, 96, 1087. [148] Bui, T.-T.; Beouch, L.; Sallenave, X.; Goubard, F. Tetrahedron Lett. 2013, 54, 4277. [149] Yamamoto, M.; Tsujii, Y.; Tsuchida, A. Chem. Phys. Lett. 1989, 154, 559. [150] Shida, T.; Nosaka, Y.; Kato, T. J. Phys. Chem. 1978, 82, 695. [151] Housecroft, C. E.; Sharpe, A. G., Anorganische Chemie, 2nd; Pearson Education Deutschland: München, 2006. [152] Curcio, J. A.; Petty, C. C. J. Opt. Soc. Am. 1951, 41, 302. [153] Büning-Pfaue, H. Food Chemistry 2003, 82, 2nd International Workshop on Water in Foods, 107. [154] Hankache, J.; Niemi, M.; Lemmetyinen, H.; Wenger, O. S. J. Phys. Chem. A 2012, 116, 8159. [155] Gutmann, V. Coord. Chem. Rev. 1976, 18, 225. [156] Reichardt, C. Chem. Rev. 1994, 94, 2319. [157] Stahl, R.; Lambert, C.; Kaiser, C.; Wortmann, R.; Jakober, R. Chem. Eur. J. 2006, 12, 2358. [158] Hudson, Z. M.; Wang, S. Acc. Chem. Res. 2009, 42, 1584. [159] Steeger, M.; Lambert, C. Chem. Eur. J. 2012, 18, 11937. [160] Zhang, Z.; Edkins, R. M.; Nitsch, J.; Fucke, K.; Eichhorn, A.; Steffen, A.; Wang, Y.; Marder, T. B. Chem. Eur. J. 2015, 21, 177. [161] Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H. Chemical Reviews 2010, 110, 6595. [162] Hide, F.; DÍaz-GarcÍa, M. A.; Schwartz, B. J.; Heeger, A. J. Acc. Chem. Res. 1997, 30, 430. [163] Kaim, W.; Schulz, A. Angew. Chem. Int. Ed. 1984, 23, 615. [164] Ricks, A. B.; Solomon, G. C.; Colvin, M. T.; Scott, A. M.; Chen, K.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem. Soc. 2010, 132, 15427. [165] Guo, Z.-H.; Jin, Z.-X.; Wang, J.-Y.; Pei, J. Chem. Commun. 2014, 50, 6088.

188

VII Bibliography [166] Lippert, E. Z. Naturforsch. A 1955, 10, 541. [167] Mataga, N.; Kaifu, Y.; Koizumi, M. Bull. Chem. Soc. Jpn. 1955, 28, 690. [168] Onsager, L. J. Am. Chem. Soc. 1936, 58, 1486. [169] Suppan, P. Chem. Phys. Lett. 1983, 94, 272. [170] Xu, X.; Ye, S.; He, B.; Chen, B.; Xiang, J.; Zhou, J.; Lu, P.; Zhao, Z.; Qiu, H. Dyes and Pigments 2014, 101, 136. [171] Florian, A.; Mayoral, M. J.; Stepanenko, V.; Fernández, G. Chem. Eur. J. 2012, 18, 14957.

189

A. Abbreviations and Symbols

0 ∆GET A A1 ADPs AQ a.u. Boc bpy CCDC CI CSS1 CSS2 CT CV Cyt f D dba DCM DMAP DMF DMSO e EPR ET Et2O EtOAc EtOH eq Fc FD FNR

free energy for electron transfer acceptor phylloquinone anisotropic displacement parameters 9,10-anthraquinone arbitrary units tert-butyloxycarbonyl 2,2’-bipyridine Cambridge Crystallographic Data Centre continuous irradiation first charge-separated state second charge-separated state charge transfer cyclic voltammetry/voltammogram cytochrome f donor dibenzylideneacetone dichloromethane 4-dimethylaminopyridine dimethyl formamide dimethyl sulfoxide elementary charge electron paramagnetic resonance electron transfer diethyl ether ethyl acetate ethanol equivalent ferrocene ferredoxin ferredoxin-NADP+ reductase

V

A Abbreviations and Symbols FT HFIP HMPA HRMS IR MCET Me MeCN MeOH MLCT NADP n-BuLi Nd:YAG NEt3 NHE NIR NMR obs OEC OLED OTA OTf P P PC PCET ph Pheo PSI PSII PQ py Q Q R ref RT

VI

Fourier transform hexafluoroisopropanol (1,1,1,3,3,3-hexafluoro-2- propanol) hexamethylphosphoramide high-resolution mass spectrometry infrared metal ion-coupled electron transfer methyl acetonitrile methanol metal-ligand charge transfer nicotinamide adenine dinucleotide phosphate n-butyllithium neodymium-doped yttrium aluminium garnet triethylamine normal hydrogen electrode near infrared nuclear magnetic resonance observed oxygen evolving complex organic light-emitting diode oligotriarylamine triflate laser excitation power photosensitizer plastocyanin proton-coupled electron transfer phenyl pheophytin photosystem I photosystem II plastoquinone pool pyridine quencher quinone distance reference room temperature

A Abbreviations and Symbols SCE TAA TBAPF6 TCNB TFA THF TMS TyrZ UV Vis XAS xy

saturated calomel electrode triarylamine tetrabutylammonium hexafluorophosphate tetracyanobenzene trifluoroacetic acid tetrahydrofuran trimethylsilyl tyrosine Z ultraviolet visible X-ray absorption spectroscopy xylene

VII

B. List of Figures

I.1

II.1

Schematic structure of a system combining light-harvesting, charge separation, and fuel production. D = donor, P = photosensitizer, A = acceptor, ET = electron transfer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

Left: Reaction center of PSII. The surrounding protein structure was omitted for clarity. Electron transfer pathways are indicated by arrows.14 Right: Crystal structure of the OEC with its ligand environment.15 Reproduced with permission from14,15 © Royal Society of Chemistry and Nature Publishing Group. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6

II.2

Simplified Jablonski diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

II.3

Potential energy wells for exergonic electron transfer in D–A including inner and outer sphere reorganization. . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

II.4

Depiction of the normal and inverted region as proposed by Marcus. . 16

II.5

Schematic diagram for electron-exchange triplet energy transfer according to Dexter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

II.6

Schematic depiction of possible reaction pathways leading to the accumulation of charge in a DPA triad. Undesired reaction pathways are depicted with dashed arrows. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

II.7

P–PBDCI–P . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

II.8

Trinuclear Ru–Ir–Ru (left) and Ru–Rh–Ru (right) complexes. . . . . . . . . . . 22

II.9

Ruthenium Photocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

II.10

TiO2–Ru–OTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

II.11

Absorption (solid trace) and emission (dashed trace) spectra of [Ru(bpy)3](PF6)2 on the left and [Re(bpy)(CO)3(py)](OTf) on the right in MeCN at RT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

II.12

Latimer diagrams for the metal photosensitizers. . . . . . . . . . . . . . . . . . . . . . 26

III.1

Molecular structure of the OTA donor. ADPs: 50% thermal ellipsoid probability. Carbon atoms: black; hydrogen atom: grey; nitrogen atoms: blue; oxygen atoms: red. (Symmetry operation used to generate equivalent atoms: 0.5-x, 0.5-y.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

IX

B List of Figures III.2

IR spectra of the investigated complexes Re–ref (black), Re2–TAA (blue), and Re2–OTA (cyan). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

III.3

UV-Vis absorption spectra of the investigated complexes Re–ref (black), Re2–TAA (blue), and Re2–OTA (cyan). . . . . . . . . . . . . . . . . . . . . . . . . 35

III.4

Cyclic voltammograms of Re2–OTA (cyan), Re2–TAA (blue), and Re– ref (black) in de-oxygenated MeCN solutions referenced to Fc+/0. . . . . . . 36

III.5

Luminescence spectra of the investigated complexes (Re2–TAA in blue and Re2–OTA in cyan) measured on de-oxygenated 10−5 M MeCN solutions after excitation at 355 nm. All spectra were corrected for differences in extinction at 355 nm and normalized to the emission of Re–ref (black). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

III.6

Steady-state luminescence spectra at of Re–ref (a), in an EtOH/MeOH 4:1) and OTA–ref (b), in MeTHF) at 77 K. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

III.7

Luminescence decays of Re–ref (black), Re2–TAA (blue), and Re2–OTA (cyan) in de-oxygenated MeCN solutions (10−5 M) after excitation at 335 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

III.8

Top: UV-Vis absorption spectrum of a 10−5 M MeCN solution of Re2– TAA (black) and its stepwise oxidation to Re2–TAA+ (blue) upon addition of increasing amounts of Cu(ClO4)2; bottom: transient absorption spectrum after 355 nm excitation of a 35 µM de-oxygenated MeCN solution of Re2–TAA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

III.9

Decay of transient absorption signals at 360 nm (blue), 490 nm (cyan) and 770 nm (orange). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

III.10 Top: stepwise chemical oxidation of a 10−5 M MeCN solution of Re2– OTA (black) to Re2–OTA+ (blue) and Re2–OTA2+ (cyan); bottom: transient absorption spectrum of a de-oxygenated 10−5 M MeCN solution of Re2–OTA directly after 355 nm excitation in a 50 ns time window. . . . 44 III.11 Transient absorption decays of the photoproduct of Re2–OTA at 420 nm and 790 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 III.12 a) Transient absorption spectrum after 355 nm excitation of a deoxygenated 10−5 M MeCN solution of Re2–OTA with an excess of TCNB; b) Corresponding decays at different wavelengths. . . . . . . . . . . . . 47 III.13 Intensity-dependent transient absorption spectra of a de-oxygenated 10−5 M MeCN solution of Re2–OTA. Cyan: 3.8 mJ, blue: 1.1 mJ, and the difference spectrum in black. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

X

B List of Figures III.14 Dependence of the signal intensity on the laser power. Left: linear dependency due to formation of (Re+)(Re0)–OTA+ in a single-photon √ P because no process. Right: non-linear behavior of intensity vs. two-photon process takes place. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 III.15 Intensity-dependent transient absorption spectra obtained from the same solutions as in III.1.10. Cyan trace: 3.1 mJ, blue trace: 1.0 mJ, black: difference spectrum obtained by subtracting the blue trace from the cyan trace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 III.16 Emission decays of Re2–TAA+ (blue) and Re2–OTA+ (cyan) after excitation of a de-oxygenated MeCN solution at 355 nm. . . . . . . . . . . . . . . . . . . 50 III.17 Near infrared spectra of Re–ref (black) and the new triad AQ–Re–OTA (blue). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 III.18 Optical absorption (left) and steady-state emission spectrum (right) of 10−5 M MeCN solutions of AQ–Re–OTA (blue) and Re–ref (black). The emission spectra were recorded after excitation of the deoxygenated complexes at 355 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 III.19 Cyclic voltammogram of AQ–Re–OTA (blue) as a 1 mM solution in MeCN. The waves at −0.5 V correspond to the internal reference Me10Fc. Re–ref (black) is depicted for direct comparison. . . . . . . . . . . . . . . 55 III.20 UV/Vis absorption spectra of 10−5 M MeCN solutions of pure OTA– ref (black), after addition of an excess of chloroacetic acid (blue), and triflic acid (cyan). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 III.21 Stepwise chemical oxidation of a 10−5 M MeCN solution of AQ–Re– OTA with Cu(ClO4)2. Blue trace = OTA+, cyan trace = OTA2+. . . . . . . . . . 58 III.22 Luminescence decays of 10−5 M MeCN solutions of pure AQ–Re–OTA (black), with 0.2 M benzoic acid (blue), and 0.2 M chloroacetic acid (cyan). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 III.23 Transient absorption spectra of de-oxygenated MeCN solutions of pure AQ–Re–OTA (black), with 0.2 M benzoic acid (blue), and 0.2 M chloroacetic acid (cyan) after 355 nm excitation. . . . . . . . . . . . . . . . . . . . . . . . 59 III.24 Transient absorption decays of photoproducts generated after excitation with a 10 ns laser pulse at 355 nm of a de-oxygenated MeCN solution of AQ–Re–OTA (a), with 0.2 M benzoic acid (b), and 0.2 M chloroacetic acid (c). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

XI

B List of Figures III.25 Transient absorption spectra of a 10−5 M MeCN solution of AQ–Re– OTA with 0.2 M chloroacetic acid at 3.4 mJ (cyan trace) and 1.1 mJ (blue trace) laser power. The blue spectrum was multiplied by a factor of 3.09 and subtracted from the cyan spectrum to give the difference spectrum in black. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 III.26 Synthetic route for preparation of Ru–OTA: a) Na2CO3, Pd(PPh3)4, H2O/EtOH/toluene, 90 °C, over night; b) ICl, DCM, 0 °C/RT, over night; c) NaOtBu, Pd(dba)2, (HPtBu3)BF4, toluene, 110 °C, 24 h; e) Ru(bpy)2Cl2, CHCl3/EtOH, reflux, 21 h. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 III.27 UV-Vis absorption (solid line) and steady state emission spectra (dashed line) of the triad (blue), dyad (cyan), and Ru–ref (black). 10−5 M MeCN solutions were used for the absorption measurements. Steady state emission was measured of de-oxygenated MeCN solutions, the spectra were corrected for differences in the absorption at 450 nm, and normalized to the Ru–ref emission. . . . . . . . . . . . . . . . . . . . . . . 67 III.28 Cyclic voltammograms of Ru–ref (black), Ru–OTA (blue) and AQ–Ru– OTA (cyan) in MeCN. The waves marked by asterisks correspond to the internal references Fc+/0 for Ru–ref and Me10Fc for the triad. . . . . . . 68 III.29 Energy level diagram for AQ–Ru–OTA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 III.30 Normalized emission decays of Ru–ref (black), Ru–OTA (blue), and AQ–Ru–OTA (cyan). The decays were recorded on 10−5 M deoxygenated MeCN solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 III.31 Difference spectra for the chemical oxidation of OTA to OTA+ (blue) and OTA2+ (cyan) obtained by addition of incremental amounts of Cu(ClO4)2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 III.32 Transient absorption spectrum (left) and corresponding transient absorption kinetics (right) of the Ru–OTA dyad in a de-oxygenated 10−5 M MeCN solution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 III.33 Left: Power-dependent transient absorption spectra of Ru–OTA in presence of an excess of TCNB after excitation at 532 nm. Blue: 10 mJ; cyan: 75 mJ. Right: Decays of the transient absorption signals at 378 (blue), 459 (cyan), and 795 nm (orange). All spectra were recorded on de-oxygenated MeCN solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 III.34 Top: Difference Spectra of OTA+ (grey trace) and OTA2+ (black trace) obtained by chemical oxidation with Cu(ClO4)2. Bottom: Difference spectrum of the high- and low-energy spectra as shown in Figure III.33 after normalization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

XII

B List of Figures III.35 Left: Transient absorption spectra after excitation of a 10−5 M MeCN solution of AQ–Ru–OTA at 532 nm. High and low-energy spectra are shown with cyan (58 mJ) and blue (2.1 mJ) traces. Difference spectrum after normalization at the isosbestic points and subtraction of the blue from the cyan spectrum (orange trace). Right: Transient absorption decays of the signals at 415, 565, and 685 nm. . . . . . . . . . . . . . . . . . . . . . . . . . 75 III.36 Transient absorption spectra of de-oxygenated 2.5 · 10−5 M MeCN solutions of AQ–Ru–OTA with 400 eq of Sc3+, Eu3+, Mg2+, Ca2+, and Ba2+. For better visualization the red and green spectra were multiplied by a factor of 5. The spectrum of the pure triad is shown as reference (black). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 III.37 Decays of the transient photoproduct in presence of various metal ions. a) Ba2+, b) Ca2+, c) Mg2+, d) Eu3+, and e) Sc3+. . . . . . . . . . . . . . . . . . . . 80 III.38 Emission decays of AQ–Ru–OTA with increasing Sc3+ concentrations. Right: Plot of kET vs c(Sc3+). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 III.39 Left: Transient absorption spectra after excitation of AQ–Ru–OTA at 532 nm in de-oxygenated MeCN with increasing amounts of Sc3+ (0.5400 eq). Right: Ground state absorption spectra of the sample solutions. The color code is the same as on the left and the spectrum in pure MeCN is depicted in black for comparison. . . . . . . . . . . . . . . . . . . . . . 81 III.40 Kinetic absorption decays of the band at 420 nm (Figure III.39) in presence of various Sc3+ concentrations in de-oxygenated MeCN solution. . 82 III.41 Left: Transient absorption spectra at high (80 mJ, cyan) and low (10 mJ, blue) pulse energies after excitation of de-oxygenated MeCN solutions at 532 nm. Difference spectrum after normalization at 478 and 1285 nm and subtraction of the low from the high energy spectrum. Right: Kinetic absorption decays of the photoproduct at different wavelengths. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 III.42 Ground-state absorption spectra of AQ–Ru–OTA in pure MeCN (black trace) and in presence of 1000 eq of diphenylthiourea (grey trace). 85 III.43 Transient absorption spectra of AQ–Ru–OTA in presence of 1000 eq diphenylthiourea after excitation at 532 nm. Top: high energy spectrum = cyan (56-66 mJ), low energy spectrum = blue (11-12 mJ). Bottom: Difference spectrum obtained by subtraction of the low from the high energy spectrum after normalization to the isosbestic points at 478 and 1285 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

XIII

B List of Figures III.44 Absorption decays of the bands at 418, 538, 680, and 780 nm recorded after excitation of the de-oxygenated, diphenylthiourea containing MeCN solution with a 532 nm laser pulse (56 mJ). . . . . . . . . . . . . . . . . . . . . 87 III.45 Left: Absorption spectra of [Fe(phen)3]2+ with increasing concentrations. Right: Calibration points and fit for the determination of the molar extinction coefficient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 III.46 Left: Absorption spectra recorded after increasing irradiation times. Right: Amount of Fe2+ (mol) vs the irradiation period. . . . . . . . . . . . . . . . . 92 III.47 UV-Vis absorption spectra after five successive measurements. a) Ru– OTA + O2; b) Ru–OTA + N2; c) AQ–Ru–OTA + O2; d) AQ–Ru–OTA + N2. Complex concentrations were 10−5 M with 0.02 M Sc(OTf)3. . . . . . . . 95 III.48 UV-Vis absorption spectra of aerated MeCN solutions of a) Ru–OTA and b) AQ–Ru–OTA after 15 min in the dark (grey traces) and five successive measurements (black traces). Complex concentrations were 10−5 M with 0.02 M Sc(OTf)3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 III.49 Changes in absorption upon continuous irradiation of aerated solutions of a) Ru–OTA (blue trace: 10 min; cyan trace: 3 h) and b) AQ– Ru–OTA (cyan trace: 30 min) in MeCN. λexc = 450 nm. . . . . . . . . . . . . . . . . 96 III.50 Top: Normalized spectra of Ru–OTA (blue) and AQ–Ru–OTA (cyan) in aerated MeCN solutions. Chemically generated OTA2+ spectrum in black. Bottom: Difference spectra of the dyad (blue) and triad (cyan) after subtraction of the OTA2+ spectral signatures. . . . . . . . . . . . . . . . . . . . . 98 III.51 Continuous irradiation of aerated MeCN solutions of Ru–OTA in presence of an excess of Sc3+. Excitation occurred at a) 430 nm and b) 500 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 III.52 Continuous irradiation of aerated MeCN solutions of Ru–OTA (a) and AQ–Ru–OTA (b) at 450 nm in absence of Sc3+. Cyan traces: after 3 h. . . . 101 III.53 Difference spectra recorded on aerated MeCN solutions kept in the dark. a) Ru–OTA + Sc3+ (blue trace = 10 min, cyan trace = 3 h), b) AQ–Ru–OTA + Sc3+ (cyan trace =30 min), c) Ru–OTA without Sc3+ (cyan trace = 3 h), and d) AQ–Ru–OTA without Sc3+ (cyan trace = 3 h). . 102 III.54 Difference spectra after continuous irradiation of de-oxygenated MeCN solutions of Ru–OTA (a)) and AQ–Ru–OTA (b)) in presence of 0.02 M Sc3+ at 450 nm. a) blue trace = 9 h. b) blue trace = 2 h, cyan trace = 9 h. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

XIV

B List of Figures III.55 Top: Difference spectra after 9 h of irradiation (black), of 47% OTA+ (blue), and 23% OTA2+ (cyan). Bottom: Difference spectrum after subtraction of the oxidized OTA contributions. . . . . . . . . . . . . . . . . . . . . . . . 104 III.56 Difference spectra after continuous irradiation of de-oxygenated MeCN solutions of Ru–OTA (a) and AQ–Ru–OTA (b) in absence of Sc3+ at 450 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 III.57 Difference spectra recorded on de-oxygenated MeCN solutions kept in the dark for the duration of 9 h. a) Ru–OTA + Sc3+, b) AQ–Ru–OTA + Sc3+, c) Ru–OTA without Sc3+, and d) AQ–Ru–OTA without Sc3+. The cyan traces correspond to the spectra recorded after 9 h in the dark.106 III.58 Absorption spectra of AQ–Ru–OTA in pure MeCN (black trace) and of the de-oxygenated, irradiated solution after addition of NEt3. . . . . . . . 107 III.59 Cyclic voltammograms of Ru–ref (black trace) and AQ–Ru–OTA–Ru– AQ (blue trace) in Ar-purged MeCN solutions. The internal references Fc+/0 for Ru–ref and Me10Fc for the pentad are marked by asterisks. . . . 110 III.60 Absorption (solid trace) and steady-state emission (dashed trace) spectra of AQ–Ru–OTA–Ru–AQ in blue. The spectra of Ru–ref are shown for comparison (black traces). Complexes were measured in de-oxygenated 10−5 M MeCN solutions. Emission spectra were corrected for differences in absorption at 450 nm and normalized to the emission of Ru–ref. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 III.61 Absorption spectra recorded on de-oxygenated, 10−5 M solutions of AQ–Ru–OTA–Ru–AQ in various solvents. MeCN = black, MeCH/H2O = blue, and HFIP = cyan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 III.62 Emission decays of AQ–Ru–OTA–Ru–AQ in de-oxygenated MeCN (black), MeCN/H2O (blue), and HFIP (cyan). Excitation occurred at 532 nm and detection at 610 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 III.63 Top: Spectral characteristics of OTA+ (blue) and OTA2+ (cyan) after chemical oxidation in MeCN. Bottom: Transient absorption spectra of AQ–Ru–OTA–Ru–AQ in MeCN (black), MeCN/H2O (blue), and HFIP (cyan). All spectra were measured of 10−5 M de-oxygenated solutions. . 114 III.64 Transient absorption decays of the photoproducts in MeCN (a), MeCN/H2O (b), and HFIP (c). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 III.65 a) MeCN, b) MeCN/H2O, and c) HFIP. Top: Intensity-dependent transient absorption spectra (cyan = high energy; blue = low energy). Bottom: Corresponding difference spectra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

XV

B List of Figures III.66 Ground state absorption spectra of 10−5 M MeCN solutions of the pure pentad (black trace) and with 2000 eq diphenylthiourea (grey trace). . . . 119 III.67 Top: Transient absorption spectra of AQ–Ru–OTA–Ru–AQ in presence of an excess of diphenylthiourea at low (10 mJ; blue trace) and high energies (60 mJ; cyan trace). Bottom: Difference absorption spectrum by subtraction of the blue from the cyan trace. The spectra were normalized to the isosbestic points (478 nm and 1285 nm) prior to subtraction. All measurements were performed in de-oxygenated MeCN solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 III.68 Stepwise chemical oxidation of AQ–Ru–OTA–Ru–AQ by addition of incremental amounts of Cu(ClO4)2. Blue trace: OTA+; cyan trace: OTA2+. The arrows mark the isosbestic points used for normalization of the transient absorption spectra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 III.69 Kinetic absorption decays of the transient absorption bands at 413, 540, and 740 nm after excitation with a laser pulse of 60 mJ in presence of an excess of diphenylthiourea. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 IV.1

Cyclic voltammograms of the six fluorophores in dry, Ar-purged THF in presence of 0.1 M TBAPF6. Left: a) TAA-short–B, b) OTA–short–B, and c) OTA–short–B2. Right: d) TAA–long–B, e) OTA–long–B, and f) OTA–long–B2. Potential sweep rates are 0.1 V/s. . . . . . . . . . . . . . . . . . . . . . 129

IV.2

Absorption (solid traces) and emission spectra (dashed traces) of TAA–short–B (black, λexc = 370 nm), OTA–short–B (blue, λexc = 330 nm), and OTA–short–B2 (cyan, λexc = 345 nm). All spectra were recorded on 10−5 M hexane solutions under ambient conditions and the emission spectra were normalized. λexc is the respective excitation wavelength. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

IV.3

Absorption (solid traces) and emission spectra (dashed traces) of TAA–long–B (black, λexc = 390 nm), OTA–long–B (blue, λexc = 345 nm), and OTA–long–B2 (cyan, λexc = 345 nm). All spectra were recorded on 10−5 M hexane solutions under ambient conditions and the emission spectra were normalized. λexc is the respective excitation wavelength. . . 132

IV.4

Absorption spectra of a) TAA–short–B, b) OTA–short–B, c) OTA– short–B2, d) TAA–long–B, e) OTA–long–B, and e) OTA–long–B2 recorded on 10−5 M solutions in hexane (black), toluene (blue), Et2O (cyan), THF (red), and DCM (orange). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

XVI

B List of Figures IV.5

IV.6 IV.7 IV.8

Emission spectra of a) TAA–short–B, b) OTA–short–B, c) OTA–short– B2, d) TAA–long–B, e) OTA–long–B, and e) OTA–long–B2 recorded on 10−5 M solutions in hexane (black), toluene (blue), Et2O (cyan), THF (red), and DCM (orange). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Lippert-Mataga plots of the Stokes shift vs. the orientation polarizability ∆ f . a) TAA–short–B, b) OTA–short–B. . . . . . . . . . . . . . . . . . . . . . . . . 135 Lippert-Mataga plot of the Stokes shift vs. the orientation polarizability ∆ f for TAA–long–B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Illustration of increased Onsager radii in TAA–short–B (left) and OTA–short–B (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

XVII

C. List of Tables

III.1 III.2 III.3 III.4 III.5 III.6 III.7

IV.1 IV.2 IV.3 IV.4

Redox potentials in MeCN. All values are given in V vs. Fc+/0. . . . . . . . . 37 Luminescence origin and triplet energies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Redox potentials in Ar-purged MeCN solutions. All values are given in V vs. Fc+/0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Used metal ions and their radii.137 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Sample preparation for the calibration curve. . . . . . . . . . . . . . . . . . . . . . . . . . 90 Extinction coefficients in M–1cm–1 for OTA+ and OTA2+ species. Percentages are given in parentheses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Gutmann acceptor numbers (AN)155 and dielectric constants ()156 of the solvents used. H2O represents the MeCN/H2O mixture. . . . . . . . . . . 117 Synthesized D–A compounds with overall yields and colors. . . . . . . . . . . 128 Redox potentials in THF given in V vs. Fc+/0. . . . . . . . . . . . . . . . . . . . . . . . . 130 Solvents and their polarity in terms of Reichhardt parameters ETN.156 . 132 Luminescence quantum yields (φ) and lifetimes (τ in ns) of the fluorophores in the different solvents. Measured in aerated solvents under ambient conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

XIX

D. Curriculum Vitae

Personal Details Name Date of Birth Birthplace

Annabell G. Bonn June 3rd, 1986 Miltenberg, Bavaria, Germany

Education 2012 - 2015

2011 - 2012

2011

2005 - 2011 1996 - 2005

University of Basel (CH) PhD thesis: “Towards Photoinduced Charge Accumulation in Purely Molecular d6 Metal Complexes with Oligo-Triarylamine Donors” Supervisor: Prof. Dr. Oliver S. Wenger Georg-August University Göttingen (D) Started PhD thesis under supervision of Prof. Dr. Oliver S. Wenger Fraunhofer Institut für Silicatforschung, Würzburg (D) Diploma thesis: “Modification of polymeric precursors for the preparation of SiC-filaments with a diameter of 100 µm for application in MMC’s” Supervisors: Dr. Andreas Nöth, Dr. Jürgen Clade Julius-Maximilians University Würzburg (D) Chemistry Studies; Degree: diploma Karl-Ernst Academic High School Amorbach (D) A-level: Chemistry, Mathematics, English, History

XXI

D Curriculum Vitae

Publications Journal Articles

First

Second

Third

Fourth

Reuter, L. G.; Bonn, A. G.; Stückl, A. C.; He, B.; Pati, P. B.; Zade, S.; Wenger, O. S. “Charge Delocalization in a Homologous Series of α,α’-Bis(dianisylamino)-Substituted Thiophene Monocations” J. Phys. Chem. A 2012, 116, 7345. Bonn, A. G.; Neuburger, M.; Wenger, O. S. “Photoinduced Electron Transfer in Rhenium(I) – Oligotriarylamine Molecules” Inorg. Chem. 2014, 53, 11075. Bonn, A. G.; Wenger, O. S. “Photoinduced Charge Accumulation in Molecular Systems” Chimia 2015, 69, 17. Bonn, A. G.; Wenger, O. S. “Charge Transfer Emission in Oligotriarylamine–Triarylborane Compounds” J. Org. Chem. 2015, 80, 4097.

Posters

First

Second

XXIV IUPAC Symposium on Photochemistry “Mixed-Valence Phenomena in Amino-Decorated Thiophene Cations” 07/2012 in Coimbra, Portugal SCS Fall Meeting “Towards Accumulative Electron Transfer in a Molecular System” 09/2013 in Lausanne, Switzerland

Talks

First

Second

XXII

PhD Community (PCC), University of Basel “Electron Transfer in the Wenger Group” 02/2013 National Centres of Competence in Research (NCCR), Molecular Systems Engineering, Lunch Meeting

D Curriculum Vitae “Working towards charge accumulation in d6 transition metal complexes” 02/2015

XXIII